@@ -170,16 +170,31 @@ There are a few cluster options
170
170
171
171
#### Greedy agent
172
172
173
- The greedy agent variant is a template modification that forces the Coder agent
174
- to transmit large metadata (size: 4K) while emitting stats. The transmission of
175
- large chunks puts extra overhead on coderd instances and agents while processing
173
+ The greedy agent variant is a template modification that makes the Coder agent
174
+ transmit large metadata (size: 4K) while reporting stats. The transmission of
175
+ large chunks puts extra overhead on coderd instances and agents when handling
176
176
and storing the data.
177
177
178
178
Use this template variant to verify limits of the cluster performance.
179
179
180
180
### Observability
181
181
182
- TODO Grafana and logs
182
+ During scale tests, operators can monitor progress using a Grafana dashboard.
183
+ Coder offers a comprehensive overview
184
+ [ dashboard] ( https://github.com/coder/coder/blob/main/scaletest/scaletest_dashboard.json )
185
+ that can seamlessly integrate into the internal Grafana deployment.
186
+
187
+ This dashboard provides insights into various aspects, including:
188
+
189
+ - Utilization of resources within the Coder control plane (CPU, memory, pods)
190
+ - Database performance metrics (CPU, memory, I/O, connections, queries)
191
+ - Coderd API performance (requests, latency, error rate)
192
+ - Resource consumption within Coder workspaces (CPU, memory, network usage)
193
+ - Internal metrics related to provisioner jobs
194
+
195
+ It is highly recommended to deploy a solution for centralized log collection and
196
+ aggregation. The presence of error logs may indicate an underscaled deployment
197
+ of Coder, necessitating action from operators.
183
198
184
199
## Autoscaling
185
200
0 commit comments