Shark IQ integration goes "Unavailable" every day

Question

Shark IQ integration goes "Unavailable" every day

AkaMrBill opened this issue 2 years ago · comments

After the "fix" was applied back in March/April, the integration worked well for a few weeks. In the last 2 months, the integration has begun going "Unavailable" every day.
"Disabling" and then "Enabling" the integration restores operation for an additional day, then it goes unavailable again.

Not being a Dev myself, I will need some guidance to help provide what you need to look into this issue.

From my logs:
Logger: homeassistant.components.sharkiq
Source: helpers/update_coordinator.py:231
Integration: Shark IQ (documentation, issues)
First occurred: June 24, 2022 at 1:14:43 PM (5 occurrences)
Last logged: June 28, 2022 at 1:30:38 PM

Authentication failed while fetching sharkiq data: Ayla Networks API authentication expired. Re-authenticate and retry.

Mark Adkins · Answer 1 · Thu Jul 21 2022 11:09:44 GMT+0800 (China Standard Time)

@AkaMrBill Have you been using an automation to reload the automation? When I upgraded my Home Assistant version last month it reset the entry id, so I had to update that in my automation that reloads the integration automatically. The refresh token is a known issue that we're tracking in #5 and unfortunately have not had a chance to tackle yet.

A temporary workaround people (including myself) have been using is an automation like this. Others have also had success making NodeRed automations to do essentially the same thing as the automation YAML below:

alias: Watcher - Shark Availability
description: ''
trigger:
  - platform: state
    entity_id: vacuum.shark
    for:
      hours: 0
      minutes: 2
      seconds: 0
    to: unavailable
condition: []
action:
  - service: homeassistant.reload_config_entry
    data:
      entry_id: <SOME_HASH_STRING>
mode: single

Put in the entity_id of your vacuum under the trigger, and the entry_id of your integration under the action data. You can get the entry id by opening up dev tools to the network tab and navigating to your devices & services page where it lists out all of your integrations. Click the context menu for the SharkIQ integration (with dev tools still open to the network tab) and click "Reload". The XHR request that shows up in dev tools should look something like this:

<YOUR_HASS_URL>/api/config/config_entries/entry/<SOME_HASH_STRING>/reload

That SOME_HASH_STRING is your entry_id. If this gets your integration working again, then we know for sure that you're hitting the refresh token issue. If thats the case please close out this issue so we can direct all discussion to #5. I'll also post this automation in that issue for reference until a fix is merged.

AkaMrBill · Answer 2 · Fri Jul 22 2022 04:48:05 GMT+0800 (China Standard Time)

Mark and Jeff, THANKS for this detailed reply! Currently I do not use Automation YAML, I use Node Red. That said, by following your detailed instructions I was able to gather my entry_id and swap it into the the automation I had set up in Node Red. It was indeed different from the old one. My initial test DID NOT FAIL as it previously had! I believe I am now up and running again! I want to let this run for a couple days to see if it is indeed resolved. If my automation does work again, I will close out the issue. Thanks again! Bill K.

…

On 7/20/2022 9:09 PM, Mark Adkins wrote: @AkaMrBill <https://github.com/AkaMrBill> Have you been using an automation to reload the automation? When I upgraded my Home Assistant version last month it reset the entry id, so I had to update that in my automation that reloads the integration automatically. The refresh token is a known issue that we're tracking in #5 <#5> and unfortunately have not had a chance to tackle yet. A temporary workaround people (including myself) have been using is an automation like this. Others have also had success making NodeRed automations to do essentially the same thing as the automation YAML below: |alias: Watcher - Shark Availability description: '' trigger: - platform: state entity_id: vacuum.shark for: hours: 0 minutes: 2 seconds: 0 to: unavailable condition: [] action: - service: homeassistant.reload_config_entry data: entry_id: <SOME_HASH_STRING> mode: single | Put in the |entity_id| of your vacuum under the trigger, and the |entry_id| of your integration under the action data. You can get the entry id by opening up dev tools to the network tab and navigating to your devices & services page where it lists out all of your integrations. Click the context menu for the SharkIQ integration (with dev tools still open to the network tab) and click "Reload". The XHR request that shows up in dev tools should look something like this: |<YOUR_HASS_URL>/api/config/config_entries/entry/<SOME_HASH_STRING>/reload | That |SOME_HASH_STRING| is your |entry_id|. If this gets your integration working again, then we know for sure that you're hitting the refresh token issue. If thats the case please close out this issue so we can direct all discussion to #5 <#5>. I'll also post this automation in that issue for reference until a fix is merged. — Reply to this email directly, view it on GitHub <#23 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AH2MTGKOC5V76XXAHLQ4IADVVC5QHANCNFSM52GCZJXA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

AkaMrBill · Answer 3 · Fri Jul 22 2022 07:00:16 GMT+0800 (China Standard Time)

Mark and Jeff, Update: NOPE! Using my process in Node Red did not restore functionality. So, now I will try your approach with a automation YAML script. I'll update you with what I find. BK

…

On 7/21/2022 2:48 PM, Bill Kendall wrote: Mark and Jeff, THANKS for this detailed reply! Currently I do not use Automation YAML, I use Node Red. That said, by following your detailed instructions I was able to gather my entry_id and swap it into the the automation I had set up in Node Red. It was indeed different from the old one. My initial test DID NOT FAIL as it previously had! I believe I am now up and running again! I want to let this run for a couple days to see if it is indeed resolved. If my automation does work again, I will close out the issue. Thanks again! Bill K. On 7/20/2022 9:09 PM, Mark Adkins wrote: > > @AkaMrBill <https://github.com/AkaMrBill> Have you been using an > automation to reload the automation? When I upgraded my Home > Assistant version last month it reset the entry id, so I had to > update that in my automation that reloads the integration > automatically. The refresh token is a known issue that we're tracking > in #5 <#5> and > unfortunately have not had a chance to tackle yet. > > A temporary workaround people (including myself) have been using is > an automation like this. Others have also had success making NodeRed > automations to do essentially the same thing as the automation YAML > below: > > |alias: Watcher - Shark Availability description: '' trigger: - > platform: state entity_id: vacuum.shark for: hours: 0 minutes: 2 > seconds: 0 to: unavailable condition: [] action: - service: > homeassistant.reload_config_entry data: entry_id: <SOME_HASH_STRING> > mode: single | > > Put in the |entity_id| of your vacuum under the trigger, and the > |entry_id| of your integration under the action data. You can get the > entry id by opening up dev tools to the network tab and navigating to > your devices & services page where it lists out all of your > integrations. Click the context menu for the SharkIQ integration > (with dev tools still open to the network tab) and click "Reload". > The XHR request that shows up in dev tools should look something like > this: > > |<YOUR_HASS_URL>/api/config/config_entries/entry/<SOME_HASH_STRING>/reload > | > > That |SOME_HASH_STRING| is your |entry_id|. If this gets your > integration working again, then we know for sure that you're hitting > the refresh token issue. If thats the case please close out this > issue so we can direct all discussion to #5 > <#5>. I'll also post this > automation in that issue for reference until a fix is merged. > > — > Reply to this email directly, view it on GitHub > <#23 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/AH2MTGKOC5V76XXAHLQ4IADVVC5QHANCNFSM52GCZJXA>. > You are receiving this because you were mentioned.Message ID: > ***@***.***> >

AkaMrBill · Answer 4 · Sat Jul 23 2022 23:21:30 GMT+0800 (China Standard Time)

This fix appears to have solved my issue! I will close out the issue. Bill Kendall

…

On 7/20/2022 9:09 PM, Mark Adkins wrote: @AkaMrBill <https://github.com/AkaMrBill> Have you been using an automation to reload the automation? When I upgraded my Home Assistant version last month it reset the entry id, so I had to update that in my automation that reloads the integration automatically. The refresh token is a known issue that we're tracking in #5 <#5> and unfortunately have not had a chance to tackle yet. A temporary workaround people (including myself) have been using is an automation like this. Others have also had success making NodeRed automations to do essentially the same thing as the automation YAML below: |alias: Watcher - Shark Availability description: '' trigger: - platform: state entity_id: vacuum.shark for: hours: 0 minutes: 2 seconds: 0 to: unavailable condition: [] action: - service: homeassistant.reload_config_entry data: entry_id: <SOME_HASH_STRING> mode: single | Put in the |entity_id| of your vacuum under the trigger, and the |entry_id| of your integration under the action data. You can get the entry id by opening up dev tools to the network tab and navigating to your devices & services page where it lists out all of your integrations. Click the context menu for the SharkIQ integration (with dev tools still open to the network tab) and click "Reload". The XHR request that shows up in dev tools should look something like this: |<YOUR_HASS_URL>/api/config/config_entries/entry/<SOME_HASH_STRING>/reload | That |SOME_HASH_STRING| is your |entry_id|. If this gets your integration working again, then we know for sure that you're hitting the refresh token issue. If thats the case please close out this issue so we can direct all discussion to #5 <#5>. I'll also post this automation in that issue for reference until a fix is merged. — Reply to this email directly, view it on GitHub <#23 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AH2MTGKOC5V76XXAHLQ4IADVVC5QHANCNFSM52GCZJXA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

AkaMrBill · Answer 5 · Sat Jul 23 2022 23:25:39 GMT+0800 (China Standard Time)

These instructions are EXACTLY what I needed. This has resolved the issue for me.

Mark Adkins · Answer 6 · Sun Jul 24 2022 00:07:16 GMT+0800 (China Standard Time)

Thanks @AkaMrBill!

I'm hoping to have a chance to finally look at the auth token issue this weekend. So hopefully an automation like this won't be necessary sooner rather than later.