Unable to action JSM alerts from Teams channels

Hi @Vipeen Sekharan ,

Adding to what @Jovânio Junior already covered, a few extra angles worth checking since this started happening "recently" and after years of it working fine:

Who originally created the integration?

In JSM-Teams ChatOps, the actions you fire from Teams are actually executed on behalf of the integration owner (the API key owner behind the scenes), not the user clicking the button. If that person left the company, was deactivated, or lost their JSM Operations permissions recently, every action in Teams will silently fail with exactly the kind of vague "Couldn't run the command" message you're seeing. Alerts keep flowing in because that direction uses a different auth. Worth confirming the owner of the integration in JSM Operations > Settings > Integrations is still active and has the right role.

Microsoft Teams "New Teams" client migration.

Microsoft has been pushing everyone to the new Teams client, and older bot behaviors don't always survive the switch cleanly, especially adaptive card actions like these. If some users in your org are on Classic Teams and others on New Teams, and both fail, that's not it, but worth ruling out.

Atlassian ChatOps bot app version in Teams.

In the Teams admin center, check whether the Atlassian bot app is on the latest version and whether app permission policies changed recently. Sometimes an org-wide policy update in Teams silently blocks bot commands even though the bot still shows as installed.

Opsgenie JSM Operations migration.

If this integration was originally set up back in the Opsgenie days and just carried over, there've been a lot of moving parts under the hood as Atlassian sunsets Opsgenie into JSM Operations. Some old ChatOps integrations need to be recreated rather than migrated. Might be worth setting up a fresh integration from scratch in a test channel and seeing if it behaves differently than the old one.

If none of the above surface anything, definitely open the ticket with Atlassian as @Jovânio Junior  suggested. Include the exact timestamp of a failed action, the correlation ID if you can grab one from the bot response, and both site URLs. That helps them trace the exact ChatOps request on their backend.

Hope you find it!

Hi Vipeen - Welcome to the Atlassian Community!

From what you described, the basic JSM ↔ Teams integration still seems healthy, because the alerts are reaching Teams and JSM correctly.

The failure appears to be specifically in the interactive ChatOps actions — acknowledge, add note, close, etc.

A few details point in that direction:

notifications are still being delivered;
the same actions work directly in the JSM interface;
the problem affects both Sandbox and Production;
reconnecting the Teams integration temporarily restored the functionality;
you are already authenticated through the Atlassian ChatOps bot.

Because of that, I would not focus first on the alert configuration itself. I would investigate the authentication/session used by ChatOps to execute commands back against JSM.

I would try:

Sign out/re-authenticate the affected user through the Atlassian ChatOps bot.
Confirm the user still has the required JSM/Operations permissions to acknowledge, add notes and close alerts.
Test the same action with an org/site admin account to rule out a user-specific permission issue.
Check whether the issue occurs in a newly created Teams channel as well.
If possible, capture the exact timestamp of a failed action and open an Atlassian Support ticket.

Since reconnecting the integration only fixed it briefly and the same behavior is occurring in both Sandbox and Production, this may be a regression or backend issue rather than a local configuration problem.

I would definitely include the timestamps, both site URLs and the screenshots when raising the support ticket so Atlassian can trace the ChatOps request on their side.

Hope this helps!