Skip to content
Latchkey

Octopus "The Tentacle is offline / unavailable" in CI

Octopus tried to communicate with the Tentacle agent on a target and could not. A listening Tentacle is not reachable on its port, or a polling Tentacle has not called home, so the machine is treated as offline and the deployment to it fails.

What this error means

The deploy fails with "The Tentacle web-01 is offline or could not be contacted" or a connection timeout to the Tentacle during a health check or deployment.

octo
Deploying to web-01...
The Tentacle 'web-01' is offline or unavailable and could not be contacted.

Common causes

The Tentacle service is stopped or unreachable

A listening Tentacle is down or blocked on its port, or a polling Tentacle stopped connecting to the server, so it cannot be contacted.

A network or thumbprint mismatch

A firewall change, DNS change, or a certificate thumbprint mismatch between server and Tentacle prevents the trusted connection.

How to fix it

Restore and verify the Tentacle connection

  1. Confirm the Tentacle service is running on the target.
  2. Check the port and firewall for listening Tentacles, or connectivity to the server for polling ones.
  3. Run a health check so the target returns to Healthy before deploying.
Terminal
octo health-check --environment "Production" \
  --server "$OCTOPUS_URL" --apiKey "$OCTOPUS_API_KEY"

Fix a thumbprint or trust mismatch

If the connection fails on trust, re-register the Tentacle so the server and target thumbprints match.

How to prevent it

  • Monitor Tentacle health and alert on offline targets.
  • Keep firewall and DNS rules stable for listening Tentacles.
  • Run a health check before deploys so offline targets surface early.

Related guides

Tired of flaky CI? Latchkey auto-heals failed jobs and retries them for you. Start free →