[Down Time] Card throws 'maximum retries exceeded' error
anuraghazra opened this issue ยท comments
Hi everyone, I see that all personal access tokens are currently failing, thus you will see the "Maximum retries exceeded" error.
I'm looking into the issue to figure out what caused it.
Ideally, it should be fixed in an hour when the PATs get rest.
Details
Public instance information
Explanation
Hey everyone, sorry for the recent downtimes.
For the sake of transparency, we get a huge amount of requests per month (it's kind of mind boggling for a side project which I built just for fun). Here's the vercel dashboard statistics:
As you can see even with 73% of the responses being cached by vercel the amount of live
requests are huge, and the bottleneck here is Personal Access Tokens
Each PAT have 5000 points, and based on the complexity of the GraphQL query & resource limitations set by GitHub, it's actually very very low considering that amount of requests.
To circumvent the issue, we currently have 7 PATs (thats 35k points) which sometimes get exhausted.
Thanks to @rickstaa we are going to add few more probably bumping it up to 10 or 12, hopefully that could help alleviate the downtimes.
Note: That these downtimes are temporary, and only lasts for 1hour, since PATs regenerate after 1 hour
So whats the idea way?
#1471 (comment)
Deploy on your own vercel instance. WHY?
It's easy to do, free & reliable.
- It reduces loads on our vercel instance.
- Your own PAT will give you more concrete stats including private stats
- It ensures that there won't be any downtime on your profile stats because you are the only one who is gonna hit your URL
Originally posted by @anuraghazra in #2130 (comment)
@anuraghazra hi, the access is still down. Do we have an expected time to fix it?
Now it's fine. Thanks.
Here still out, i need do something?
Same here, any known fix to this?
Never imagined that my profile would be that much dependent on the GitHub stats , ๐.
Hope that will be done soon, nice work man.
Hi everyone.
Sorry for the inconvenience.
I will highly recommend deploying github-readme-stats on your own vercel instance to fix this issue, while I try to find a fix. Currently I'm unable to find any root cause.
Consider deploying on your own fork. follow the guide -> https://github.com/anuraghazra/github-readme-stats#deploy-on-your-own-vercel-instance
Youtube video tutorial: https://www.youtube.com/watch?v=n6d4KHSKqGk&feature=youtu.be&t=107
Also after forking make sure to open vercel.json and change the maxDuration option to 10.
Thanks a lot! mine is already bug fixed
Thanks a lot! mine is already bug fixed
Yes, it's happening intermittently for some people, and for some it's not.
Its happening again.
He is working on the fix, till then, he said to run it on our own vercel instance
Hi everyone.
Sorry for the inconvenience. I will highly recommend deploying github-readme-stats on your own vercel instance to fix this issue, while I try to find a fix. Currently I'm unable to find any root cause.
Consider deploying on your own fork. follow the guide -> https://github.com/anuraghazra/github-readme-stats#deploy-on-your-own-vercel-instance
Youtube video tutorial: https://www.youtube.com/watch?v=n6d4KHSKqGk&feature=youtu.be&t=107
Also after forking make sure to open vercel.json and change the maxDuration option to 10.
Added 4 new PATs, let's observe the downtime.
For some reason the PATs are weirdly depleting super fast. not sure why.
ALSO: Note that deploying on your own instance is ALWAYS a better option than relaying on ours
- It reduces loads on our vercel instance, we get almost 3-4Million requests per month which is hard to scale up.
- Your own PAT will give you more concrete stats including private stats
- It ensures that there won't be any downtime on your profile stats because you are the only one who is gonna hit your URL
So please I enocurage everyone to deploy on their own instance,
If you face any kind of issues while deployment I'm more than happy to help you out.
It works now !! ๐
Yeah, thanks, gotta deploy my own instance now
Yes, it's happening intermittently for some people, and for some it's not.
Right, it is failing intermittently - checked around 6 PM IST. As @anuraghazra suggested already: deploying github-readme-stats on your own vercel instance
would be a better option as of now.
Just so you know, I just ran into this problem too.
Yaah same problem here, now going forward I 'am planning to deploy this in personal vercel app :)
suggest to deploy your own instance
Happening to me
I'm facing the same issue as well
The same issue prevails here.
Hi folks, I think the PATs are again depleting (even though we have 6 PATs now)
Please consider #1471 (comment)
A friendly ping to indicate that this is issue is occurring again :|
tagging @anuraghazra to let you know about the issue again
It works well on this hour, and I have deploy on my own vercel, to take a little less stress on this repository's vercel instance :)
@anuraghazra Maybe we should add this to the bug template FAQ (i.e. https://github.com/anuraghazra/github-readme-stats/edit/master/.github/ISSUE_TEMPLATE/bug_report.md)? Also, maybe it is better to put the FAQ on top?
It looks like 9d2bfd9 fixed this issue for now! ๐ Please feel free to ping us if you experience any problems in the future. However, only comment if nobody has commented in the last three days or if you have information that could help us solve the problem. If there already is a recent comment or the label is attached, we are aware of the issue and are working on it. ๐ค Comments like "+1" and "same here" don't help anyone; they pressure the maintainer and spam their inbox. ๐ฅ
"Maximum retries exceeded" right now
@weerdy15 Thanks a lot for letting us know that this is still happening! I changed fixed to in-progress again.
In the meantime, please consider deploying your own Vercel instance to reduce server load (see github-readme-stats#deploy-on-your-own-vercel-instance). Keeping your fork up to date is just a click of a button. This process can also be automated using a GitHub action like github-repo-sync. I also added a GitHub feature request to ease the process of keeping a fork up to date (see community/community#16913).
The issue should be fixed for now as we added three new Personal Access Tokens (PATs).
The issue should be fixed for now as we added three new Personal Access Tokens (PATs).
Can I give you mine too?
@ItsAleph Thanks a lot for offering, but that is unnecessary and not wise from a security standpoint. @anuraghazra and I will create new PAT ourselves when the current PATs run out. Are you currently experiencing problems?
As mentioned in #1986, the limits are being hit again. To resolve this, the number of PATs needs to be increased. For people that are experiencing this issue, please consider deploying your Vercel instance (see https://github.com/anuraghazra/github-readme-stats#deploy-on-your-own-vercel-instance). Doing so would decrease the load on the public server. Thanks in advance ๐.
The @codeSTACKr video is outdated, the vercel process changed. Same for the image guide from https://github.com/anuraghazra/github-readme-stats#deploy-on-your-own-vercel-instance.
And by the way just forking doesn't work anymore, during vercel deployment you will get this error:
Error! The value for maxDuration must be between 1 second and 10 seconds, in order to increase this limit upgrade your plan: https://vercel.com/pricing
So one has to edit vercel.json
before if you plan to use the free tier:
{
"functions": {
"api/*.js": {
"memory": 128,
- "maxDuration": 30
+ "maxDuration": 10
}
},
"redirects": [
{
"source": "/",
"destination": "https://github.com/anuraghazra/github-readme-stats"
}
]
}
So one has to edit
vercel.json
before if you plan to use the free tier
Thanks, that fixed it! I originally got the same error but the project still got created in Vercel. Simply editing the vercel.json
file in the cloned repository and redeploying wouldn't cut it and it just kept returning Maximum retries exceeded. For anyone struggling with this just start over, fork this repository, edit the vercel.json
file as explained above and then import it into a new Vercel project - and don't forget to add your PAT_1 environment variable.
The @codeSTACKr video is outdated, the vercel process changed. Same for the image guide from https://github.com/anuraghazra/github-readme-stats#deploy-on-your-own-vercel-instance.
And by the way just forking doesn't work anymore, during vercel deployment you will get this error:
Error! The value for maxDuration must be between 1 second and 10 seconds, in order to increase this limit upgrade your plan: https://vercel.com/pricing
So one has to edit
vercel.json
before if you plan to use the free tier:{ "functions": { "api/*.js": { "memory": 128, - "maxDuration": 30 + "maxDuration": 10 } }, "redirects": [ { "source": "/", "destination": "https://github.com/anuraghazra/github-readme-stats" } ] }
@noraj You are correct the video was done before we needed to increase the maxDuration
parameter. Thanks for elaborating on that point! ๐๐ป However, we updated the text guide some time ago to include the maxDuration
step (see step 6). Do you think the documentation can still be improved?
@anuraghazra Maybe you can add the PAT I send you over Twitter? Further, maybe we can update the README a bit and add a note at the top explaining that the Vercel instance sometimes runs into GitHub query limits and point out that deploying a Vercel instance is the recommended option. Additionally, I think @4l1fe is implementing a docker container for the GRS script. If that feature is implemented, I will make a small GitHub action that can be used to update the cards periodically to give people more options.
@noraj You are correct the video was done before we needed to increase the
maxDuration
parameter. Thanks for elaborating on that point! ๐๐ป However, we updated the text guide some time ago to include themaxDuration
step (see step 6). Do you think the documentation can still be improved?
Yeah, for example there is no longer an import button on the dashboard but that instead ๐๐ป
old import button
this is what you got on new project creation now (looking on you organization / namespace repository or a small link to import git repo external to your oAuth app)
old import form
etc. the interface and workflow totally changed so the video and screenshots must all be redone from scratch unfortunately
@anuraghazra Maybe you can add the PAT I send you over Twitter? Further, maybe we can update the README a bit and add a note at the top explaining that the Vercel instance sometimes runs into GitHub query limits and point out that deploying a Vercel instance is the recommended option. Additionally, I think @4l1fe is implementing a docker container for the GRS script. If that feature is implemented, I will make a small GitHub action that can be used to update the cards periodically to give people more options.
I'm on the way...
@anuraghazra Maybe you can add the PAT I send you over Twitter? Further, maybe we can update the README a bit and add a note at the top explaining that the Vercel instance sometimes runs into GitHub query limits and point out that deploying a Vercel instance is the recommended option. Additionally, I think @4l1fe is implementing a docker container for the GRS script. If that feature is implemented, I will make a small GitHub action that can be used to update the cards periodically to give people more options.
I'm on the way...
Great! ๐ No hurry the real solution would be to add an extra PAT.
@noraj You are correct the video was done before we needed to increase the
maxDuration
parameter. Thanks for elaborating on that point! ๐๐ป However, we updated the text guide some time ago to include themaxDuration
step (see step 6). Do you think the documentation can still be improved?Yeah, for example there is no longer an import button on the dashboard but that instead ๐๐ป
old import button
this is what you got on new project creation now (looking on you organization / namespace repository or a small link to import git repo external to your oAuth app)
old import form
etc. the interface and workflow totally changed so the video and screenshots must all be redone from scratch unfortunately
Thanks for letting us know. I created #1990, so we know the problem and somebody can take it on.
So one has to edit
vercel.json
before if you plan to use the free tierThanks, that fixed it! I originally got the same error but the project still got created in Vercel. Simply editing the
vercel.json
file in the cloned repository and redeploying wouldn't cut it and it just kept returning Maximum retries exceeded. For anyone struggling with this just start over, fork this repository, edit thevercel.json
file as explained above and then import it into a new Vercel project - and don't forget to add your PAT_1 environment variable.
I needed to edit vercel.json
and only replace PAT with a newly-generated one, then re-deploy. No need to make a new project ;D
Hi, how was it requested in #1772 I'm just here to say that the Vercel limit was reached here in the last hour, more precisely between 12:00-13:00 in Brazil (GMT-3)
@RafaelSolVargas Thanks for letting us know. ๐๐ป We will add another PAT if this happens more frequently. If you want, you can deploy your own Vercel instance (see https://github.com/anuraghazra/github-readme-stats#deploy-on-your-own-vercel-instance) to prevent this from happening.
limit is being reached currently. I am aware of the guide on deploying an instance, but just letting you know.
We added an extra PAT. The issue should be fixed for now.
Its back lol
It's not really.
@anuraghazra, I think we should also increase the retry number in the retry function.
github-readme-stats/src/common/retryer.js
Line 13 in 94d0978
see #2121.
It's not really.
The BUG is back... I should be more clear lol
the bug is back to me
Posting this here as a reference:
Hey everyone, sorry for the recent downtimes.
For the sake of transparency, we get a huge amount of requests per month (it's kind of mind boggling for a side project which I built just for fun). Here's the vercel dashboard statistics:
As you can see even with 73% of the responses being cached by vercel the amount of live
requests are huge, and the bottleneck here is Personal Access Tokens
Each PAT have 5000 points, and based on the complexity of the GraphQL query & resource limitations set by GitHub, it's actually very very low considering that amount of requests.
To circumvent the issue, we currently have 7 PATs (thats 35k points) which sometimes get exhausted.
Thanks to @rickstaa we are going to add few more probably bumping it up to 10 or 12, hopefully that could help alleviate the downtimes.
Note: That these downtimes are temporary, and only lasts for 1hour, since PATs regenerate after 1 hour
So whats the idea way?
#1471 (comment)
Deploy on your own vercel instance. WHY?
It's easy to do, free & reliable.
- It reduces loads on our vercel instance.
- Your own PAT will give you more concrete stats including private stats
- It ensures that there won't be any downtime on your profile stats because you are the only one who is gonna hit your URL
Originally posted by @anuraghazra in #2130 (comment)
This bug is improved when the following PRs are merged:
Hey there, I just encountered the same error as probably most here over the last month.
Just curious: What would prevent us/you/someone from hosting this on a VPS or sth? The request load doesn't seem that bad (10 requests/s) after all. Or is there no issue with vercel at all and its just the GitHub API quota being used up?
Hey there, I just encountered the same error as most here over the last month. Just curious: What would prevent us/you/someone from hosting this on a VPS or sth? The request load doesn't seem that bad (10 requests/s). Or is there no issue with vertical at all, and it's just the GitHub API quota being used up?
@CloudyByte Thanks for letting us know and for your question. There exist Vercel Limits, but we are not close to hitting these because we using caching. The GraphQL limits currently cause the PAT error (see https://docs.github.com/en/graphql/overview/resource-limitations). Hosting on another provider will, therefore, not help.
The PRs explained in #1471 (comment) will improve the downtime. Apart from that, the best option is hosting your own Vercel instance since you can use your own GitHub quota (see https://github.com/anuraghazra/github-readme-stats#deploy-on-your-own-vercel-instance). I'm also planning on releasing a GitHub action version (see #2179). This would allow for more accurate stats, no downtime and faster loading times. I am, however, waiting on some other PRs to be merged/completed before continuing this feature.
@CloudyByte When did you encounter the error? Our downtime monitor (queries every 15 min) does not show any downtime in the last seven days.
Maybe it was down for a minimal amount of time?
Very strange. I don't have this error when I request your card at https://github-readme-stats.vercel.app/api?username=cloudybyte&show_icons=true&locale=en&theme=cobalt.
[![Anurag's GitHub stats](https://github-readme-stats.vercel.app/api?username=cloudybyte&show_icons=true&locale=en&theme=cobalt)](https://github.com/anuraghazra/github-readme-stats)
Peek.2022-10-18.17-05.mp4
@CloudyByte Thanks for letting us know and for your question. There exist Vercel Limits, but we are not close to hitting these because we using caching. The GraphQL limits currently cause the PAT error (see https://docs.github.com/en/graphql/overview/resource-limitations). Hosting on another provider will, therefore, not help.
The PRs explained in #1471 (comment) will improve the downtime. Apart from that, the best option is hosting your own Vercel instance since you can use your own GitHub quota (see https://github.com/anuraghazra/github-readme-stats#deploy-on-your-own-vercel-instance). I'm also planning on releasing a GitHub action version (see #2179). This would allow for more accurate stats, no downtime and faster loading times. I am, however, waiting on some other PRs to be merged/completed before continuing this feature.
Its really not bothering me enough to set up my own vercel instance, but I wanted to improve the experience for everyone.
Very strange. I don't have this error when I request your card at https://github-readme-stats.vercel.app/api?username=cloudybyte&show_icons=true&locale=en&theme=cobalt.
[![Anurag's GitHub stats](https://github-readme-stats.vercel.app/api?username=cloudybyte&show_icons=true&locale=en&theme=cobalt)](https://github.com/anuraghazra/github-readme-stats)
Thats really strange... Its working in your comment, but broken on my profile...
Thats probably sth to look into
@CloudyByte Thanks for letting us know and for your question. There exist Vercel Limits, but we are not close to hitting these because we using caching. The GraphQL limits currently cause the PAT error (see https://docs.github.com/en/graphql/overview/resource-limitations). Hosting on another provider will, therefore, not help.
The PRs explained in #1471 (comment) will improve the downtime. Apart from that, the best option is hosting your own Vercel instance since you can use your own GitHub quota (see https://github.com/anuraghazra/github-readme-stats#deploy-on-your-own-vercel-instance). I'm also planning on releasing a GitHub action version (see #2179). This would allow for more accurate stats, no downtime and faster loading times. I am, however, waiting on some other PRs to be merged/completed before continuing this feature.Its really not bothering me enough to set up my own vercel instance, but I wanted to improve the experience for everyone.
That's very nice of you help is always welcome! ๐
Very strange. I don't have this error when I request your card at https://github-readme-stats.vercel.app/api?username=cloudybyte&show_icons=true&locale=en&theme=cobalt.
[![Anurag's GitHub stats](https://github-readme-stats.vercel.app/api?username=cloudybyte&show_icons=true&locale=en&theme=cobalt)](https://github.com/anuraghazra/github-readme-stats)Thats really strange... Its working in your comment, but broken on my profile... Thats probably sth to look into
It could be a caching problem, but I thought I fixed it 4df8094. Maybe the header is still incorrect. Could you maybe do the following troubleshooting for me?
Steps
- Open https://github-readme-stats.vercel.app/api?username=cloudybyte&show_icons=true&locale=en&theme=cobalt in a separate browser tab to see if it works.
- Open https://github-readme-stats.vercel.app/api?username=cloudybyte&show_icons=true&locale=en&theme=cobalt in another browser to see if it works.
- Can you check if your browser stored cache (see https://developer.chrome.com/docs/devtools/storage/cache/).
2022-10-18.17-14-19.mp4
- works
- works (tested in chromium, ff is my default)
- there is apparently no local cache for
*.vercel.app
saved
That sounds a lot like a FF bug to me, if you prevented caching of errors by setting the correct response header...
- works
- works (tested in chromium, ff is my default)
- there is apparently no local cache for
*.vercel.app
savedThat sounds a lot like a FF bug to me, if you prevented caching of errors by setting the correct response header...
Thanks for the information! ๐ช๐ป I will check the caching headers later this week. What is a FF bug ๐ค?
O I think you mean Firefox! Good to know! The caching behaviour differs per browser. I will double-check the headers.
@CloudyByte I can reproduce your error in firefox. Can you maybe check the caching headers in FF for me (see video)?
Peek.2022-10-18.17-30.mp4
@CloudyByte One last thing that would help after you gave me the cache header information is to check if the problem goes away when you click the Disable Cache button. ๐๐ป For me, that solves the problem, so the no-storage header is not working on Firefox.
Mhm, maybe not. I tried requesting my profile with Chromium and FF in private mode and both are working.
The only issue is with my non-private-mode FF (firefox-developer-edition 106.0b9 (64-bit)
to be precise)
Ooh, I think I know where the issue is and why it only works when requesting directly: GitHub seems to cache the image itself when building the profile page. The error is probably cached on their side. That's why requesting it directly from the Vercel instance works.
Great detective work! Could be the issue. I will check if the GitHub CDN respects the no-store
header.
Based on my other findings, it doesn't look like GitHub is being the issue.
Based on my other findings, it doesn't look like GitHub is the issue.
So it just works if you use the GitHub CDN URL in any other browser?
Yea
Yea
It looks like a particular Vercel + Firefox error. The card shown in #1471 (comment) is coming from the Vercel cache (time 15:30). I will check at the Vercel GitHub.
Yea
I think firefox is not respecting the no-cache
header then. The private mode does not store cache
Yea
It looks like a particular Vercel + Firefox error. The card shown in #1471 (comment) is coming from the Vercel cache (time 15:30). I will check at the Vercel GitHub.
Changing the error cache header to Cache-Control: no-cache, no-store, must-revalidate
might solve the issue (see https://stackoverflow.com/a/2068407/8135687).
Can you try fetching the stats directly using this url: https://github-readme-stats.vercel.app/api?hide_rank=true&show_icons=true&include_all_commits=false&count_private=true&disable_animations=false&theme=dark&locale=en&hide_border=true&custom_title=Github Stats&username=Kentakoong
and tell us which browser u're using? (Your profile is broken on my side, but fetching directly works fine...)
![]()
mine's not showing the GH Stats and the Top Langs
ps: I'm using the img tag so I guess it's a caching issue like said in the upper comments!
@Kentakoong, now it doesn't seem to be a caching issue anymore. The PATs quota for this hour has simply been used up. To decrease this occurrence, #2121 has to be merged, and some PATs have to be replaced (see #2121). I will check back in an hour.
Can you try fetching the stats directly using this url:
https://github-readme-stats.vercel.app/api?hide_rank=true&show_icons=true&include_all_commits=false&count_private=true&disable_animations=false&theme=dark&locale=en&hide_border=true&custom_title=Github Stats&username=Kentakoong
and tell us which browser u're using? (Your profile is broken on my side, but fetching directly works fine...)
You are right! Fetching directly still works, so maybe the PATs still have quota left, and this is a cached response on the server side that is still being shown from the GitHub CDN. Very strange problem.
Your syntax markdown
[![Anurag's GitHub stats](https://github-readme-stats.vercel.app/api?hide_rank=true&show_icons=true&include_all_commits=false&count_private=true&disable_animations=false&theme=dark&locale=en&hide_border=true&custom_title=Github Stats&username=Kentakoong)](https://github.com/anuraghazra/github-readme-stats)
Your syntax with img tag
<div align="center">
<img src="https://github-readme-stats.vercel.app/api?hide_rank=true&show_icons=true&include_all_commits=false&count_private=true&disable_animations=false&theme=dark&locale=en&hide_border=true&custom_title=Github Stats&username=Kentakoong" height="150" alt="ghstats" />
</div>
Your syntax in markdown while breaking the cache with &safdsaf=adsfsaf
[![Anurag's GitHub stats](https://github-readme-stats.vercel.app/api?hide_rank=true&show_icons=true&include_all_commits=false&count_private=true&disable_animations=false&theme=dark&locale=en&hide_border=true&custom_title=Github Stats&username=Kentakoong&safdsaf=adsfsaf)](https://github.com/anuraghazra/github-readme-stats)
Your syntax with img tag while breaking the cache with &safdsaf=adsfsaf
<div align="center">
<img src="https://github-readme-stats.vercel.app/api?hide_rank=true&show_icons=true&include_all_commits=false&count_private=true&disable_animations=false&theme=dark&locale=en&hide_border=true&custom_title=Github Stats&username=Kentakoong&safdsaf=adsfsaf" height="150" alt="ghstats" />
</div>
Static image of the results above in case they update
@CloudyByte and @
Kentakoong It Looks like the problem at the current time only is present when the <IMG>
tag is used (see #1471 (comment)). This, however, is, I think, a coincidence and could as well have been the other way around.
I think the GitHub CDN could still cause this caching problem. This behaviour can be seen in action if I used the cached image URL directly:
<div align="center">
<img src="https://camo.githubusercontent.com/bdc4b7001ae4db64d42b0567e58c40584162aff3cd12f250869298dafec66d78/68747470733a2f2f6769746875622d726561646d652d73746174732e76657263656c2e6170702f6170693f686964655f72616e6b3d747275652673686f775f69636f6e733d7472756526696e636c7564655f616c6c5f636f6d6d6974733d66616c736526636f756e745f707269766174653d747275652664697361626c655f616e696d6174696f6e733d66616c7365267468656d653d6461726b266c6f63616c653d656e26686964655f626f726465723d7472756526637573746f6d5f7469746c653d476974687562c2a0537461747326757365726e616d653d4b656e74616b6f6f6e67" height="150" alt="ghstats" />
</div>
To be sure, I asked a question in the Vercel repository.
Static image in case GitHubs CDN gets updated
@CloudyByte I think I found the problem. The GitHub CDN needs the no-cache
(see https://docs.github.com/en/authentication/keeping-your-account-and-data-secure/about-anonymized-urls#an-image-that-changed-recently-is-not-updating). I updated the cache using the curl -X PURGE https://camo.githubusercontent.com/bdc4b7001ae4db64d42b0567e58c40584162aff3cd12f250869298dafec66d78/68747470733a2f2f6769746875622d726561646d652d73746174732e76657263656c2e6170702f6170693f686964655f72616e6b3d747275652673686f775f69636f6e733d7472756526696e636c7564655f616c6c5f636f6d6d6974733d66616c736526636f756e745f707269766174653d747275652664697361626c655f616e696d6174696f6e733d66616c7365267468656d653d6461726b266c6f63616c653d656e26686964655f626f726465723d7472756526637573746f6d5f7469746c653d476974687562c2a0537461747326757365726e616d653d4b656e74616b6f6f6e67
command and now your profile is working again.
Mhm, that sounds weird. Thanks for fixing it tho.
hmu if u need help or anything
Yea strange behaviour indeed the GitHub CDN should also respect no-store
header. I think 98f9045 will fix the issues. Please feel free to comment below if it occurs again.
This bug is improved when the following PRs are merged:
- fix: fix retry max-out bug #2121: Adds support for more PATs.
- feat: improve card loading speed #2124: Speeds up initial card loading.
- feat: enable multi page star fetching for private vercel instances #2159: Removes one GraphQl function call, reducing the PAT deterioration rate.
- fix: prevent totalCommitsFetch error result from being cached #2177: Makes sure that errors are not cached.
- feat: add PAT monitoring functions #2178: Should be checked since it looks like some PATs are expired.
@HADMARINE, thanks for letting us know the instance was indeed down for about 10 minutes. @anuraghazra, can you look at the issues above if you find the time? Especially #2178 and #2159. It seems that only one of our PATs is still working. ๐ค
@anuraghazra @rickstaa I'm receiving it this morning
self doing mine worked, I had to purge it though.
We are aware of the issue and actively monitoring the PATs thanks for letting us know. ๐๐ป
Thank you so much for your reply.
I followed the link in readme (https://vercel.com/new/clone?repository-url=https%3A%2F%2Fgithub.com%2Fanuraghazra%2Fgithub-readme-stats) and it didn't solve..
Do you know how I could get personal access token and put it in a env var called PAT_1?
I have no idea how I should do it..
if you have ever used git on the cmdline its basically the same thing
here`s the url to make a token: https://github.com/settings/tokens
how to add an env var to Vercel: https://vercel.com/docs/concepts/projects/environment-variables
@dginovker Thanks for your offer. I, however, have more than enough we just need to add them to Vercel. ๐๐ป But most importantly, it is best if more people deploy their own Vercel instance (see https://github.com/anuraghazra/github-readme-stats#deploy-on-your-own-vercel-instance).
@rickstaa @anuraghazra this appeared again