Deployment with Ops manager Failed

I have a sharded cluster setup with Ops manager community edition. It worked fine, however, recently it stopped working on any deployments. After checking the agent logs, I found it seems that any command is refused because of an unauthorized problem.

(Unauthorized) not authorized on local to execute command { find: “clustermanager”, filter: {}, limit: 1, singleBatch: true, sort: {}, lsid: { id: UUID(“08350143-d401-4583-91cc-75be76aa76d2”) }, $db: “local”, $readPreference: { mode: “primaryPreferred” } }

Above is sample log collected, agent credentials should be managed by opsmanager itself and we can’t change. I am not sure what’s happening here or how do I throubleshoot here?