PipelineError: Could not fetch a request token from Jira. Unable to reach host: jira.corp.adcinte...

[sentry-io]

Sentry Issue: SENTRY-167C

PipelineError: Could not fetch a request token from Jira. Unable to reach host: jira.corp.adcinternal.com

[kpujjigit]

Does Sentry cache IP addresses it resolves from DNS for Jira, leading to hits on the wrong IP address?

Context: Customer is trying to install our JIRA Server integration, Their network team found Sentry is reaching an incorrect IP address for Jira, suspecting Sentry might be caching IP addresses from DNS resolutions.

Their DomainName should resolve to their public IP (whitelisted and internally translated via NAT to a private IP for Jira). Sentry seems to be hitting a different public IP address, which is blocked by the customer as it's not authorized.

So far: I've verified they've whitelisted the correct IPs, their consumer key and key pairs are correct. In fact, the integration has worked in the past for them with the exact same credentials. Historically, every few months, the integration would stop working. They've uninstalled and reinstalled to get it working again, only this time a reinstall is not resolving and they're losing insight.

The proposed next step (by the customer) is for Sentry to execute 'test-NetConnection <Customer's_FQDN> -Port 443' to verify if the correct IP address is being resolved, enabling the customer to check their logs for confirmation. It's not a bad idea; is this something we can do?

[getsantry]

Routing to @getsentry/product-owners-settings-integrations for triage ⏲️

[ceorourke]

It looks like the TTL is 1 hour. I'm making an ops ticket to try to see if we can access their IP, I'm not sure how long it'll take for that to be picked up.