estuary / airbyte

https://airbyte.com
Other
6 stars 3 forks source link

SaaS connectors with random errors #192

Closed dyaffe closed 1 year ago

dyaffe commented 1 year ago

Problem We have a bunch of SaaS connectors that have been running for a while and are now giving random errors. We should attempt to identify the issues so that the connectors are reliable and our alerting won't be clogged -- they are below:

  1. Jira: estuary/jira/source-jira

raise HTTPError(http_error_msg, response=self)\nrequests.exceptions.HTTPError: 400 Client Error: Bad Request for url: https://estuary1.atlassian.net/rest/agile/1.0/board/1/sprint?maxResults=50"}

(generally working and getting data)

  1. Youtube: estuary/dave-test12233/source-youtube-analytics

Error while refreshing access token: 400 Client Error: Bad Request for url: https://oauth2.googleapis.com/token"}

(generally working and getting data)

  1. Zendesk Estuary-New/j/source-zendesk-chat is getting this error:

401 Client Error: Unauthorized for url: https://www.zopim.com/api/v2/triggers?limit=100"}

  1. Sentry Estuary-New/testsentry/source-sentry is getting:

HTTPError(http_error_msg, response=self)\nrequests.exceptions.HTTPError: 404 Client Error: Not Found for url: https://sentry.io/api/0/projects/estuary/jennyproject/events/?full=true"}

  1. Freshdesk fjtest/test1/source-freshdesk

HTTPError: 403 Client Error: Forbidden for url: https://fjtest.freshdesk.com/api/v2/canned_response_folders?per_page=100"}

  1. Chargebee estuary/mahdi-chargebee/source-chargebee

File \"/usr/local/lib/python3.9/site-packages/airbyte_cdk/sources/declarative/retrievers/simple_retriever.py\", line 327, in parse_response\n raise ReadException(error_message)\nairbyte_cdk.sources.declarative.exceptions.ReadException: Request to https://estuary-test.chargebee.com/api/v2/addons?sort_by%5Basc%5D=updated_at\u0026include_deleted=true\u0026updated_at%5Bbetween%5D=%5B1680307200%2C+1682899199%5D failed with status code 403 and error message The site is not enabled"}

This may just be end of a free trial.

  1. Notion

The text from Notion pages get loaded to rich_text. To get that data I currently have to swap rich_text with text in the schema specs. I think the schema that notion outputs on discovery needs to have text and rich_text properly mapped?

mdibaiee commented 1 year ago
  1. This one needs more investigation
  2. This error seems to have happened before (on 29th), but recovered. This time though it seems we are stuck in a kind of loop with this error. This might be us exhausting our rate limits. One issue here is that we do not hold the connector off from restarting if there is an error (the run_interval does not apply), only if it is successful. So in this case, this task has always errored since we applied the run_interval, so we could never give it a break. My suggestion is giving it a break manually by disabling the task for a day and then enabling it after.
  3. Strong suspicion it's a free trial ending, the error started 12 days after setting up the capture (perhaps two days after we got the trial?)
  4. This one has never worked apparently, since the beginning we have had these logs. I would first check to see if we actually have that project estuary/jennyproject
  5. Also likely to be a free trial issue. This error started appearing 19 days after they set up the capture. Freshdesk has a 21 day trial so that makes sense.
  6. Also a free trial ending
dyaffe commented 1 year ago
  1. Disabled and will re-enable tomorrow
  2. @jman0119 are you using our Zendesk account for this? We should be past a free trial since we pay. If not, want to disable it?
  3. @jman0119 please take a look
  4. Will disable
  5. Will disable
jman0119 commented 1 year ago
  1. I used another Zendesk trial account for this, will disable
  2. Disabled.
jman0119 commented 1 year ago
  1. Added Notion to the list.