diff --git a/READMES/devops/tugboat.md b/READMES/devops/tugboat.md
index bb0ab10c4e..c2446d7615 100644
--- a/READMES/devops/tugboat.md
+++ b/READMES/devops/tugboat.md
@@ -33,3 +33,23 @@ Can only update CPU and memory at a project level, not repository level.
 1. `tugboat ls 5fd3b8ee7b465711575722d5 -j | grep memory` # Get current limit
 1. `tugboat update 5fd3b8ee7b465711575722d5 memory=16384` # Set limit to 16GB
 1. `tugboat ls 5fd3b8ee7b465711575722d5 -j | grep memory` # Verify new limit
+
+## Tugboat Crisis Intervention
+
+### Overload
+
+**Symptoms**: Tugboat is slow, requests to Tugboat dashboard return 502/504 status codes, previews disappear and reappear, etc.
+
+**Diagnosis**: Tugboat might be overloaded; too many previews might be running simultaneously.
+
+**Verification**: 
+
+1. Log into the Tugboat server (`ssm-session utility tugboat auto`).
+2. Check system load and free memory (e.g. `top`).
+3. If load is incredibly high, and available memory is very low, then the Tugboat server might be dealing with too many open previews.
+
+**Remediation**:
+
+1. Close unused previews in the CMS/Pull Requests project. Target older previews and those corresponding to closed/merged PRs; these should be closed automatically, but there may be issues somewhere in the system that impair communication and cause these to remain open.
+2. Suspend older previews. This normally happens automatically (for Pull Request-based previews that haven't been touched in some period of time), but a flurry of previews might have been created inadvertently.
+3. Consider upscaling the Tugboat server or migrating to an alternative architecture.