Warm start
Three runs after a cold start, with the golden base layers kept. Seconds until every container is healthy.
13 s
median to healthy
13 to 13 s
range over warm runs
skipped
golden base layers
Cold start
One run with the golden layers, service images and build cache removed first: a full rebuild.
50 s
to healthy, full rebuild
21 s
golden base layers
45 s
span of the image builds
Per service
Image build time per service, in seconds, for each run.
| Service | cold-1 | warm-1 |
|---|---|---|
| reservation-api | 10.3 | 1.0 |
| kitchen-api | 10.9 | 1.4 |
| menu-api | 11.6 | 1.7 |
| reservation-app | 10.5 | 1.8 |
| kitchen-worker | 7.8 | 1.3 |
| floor-app | 9.5 | 1.6 |
History
Every published result in benchmarks/results. Each machine is different; compare a row with its own environment, not across rows.
| Result | Cold to healthy (s) | Warm median to healthy (s) |
|---|---|---|
| 2026-10-11 | 281 | 40 |
Environment
| Commit | ef6c45ac8d7c2ba3c02ebdd693adc6b98433a7fc |
|---|---|
| Workflow run | https://github.com/flisk-sh/flisk/actions/runs/39565225282 |
| tdk | 1.3.189 |
| tilt | v0.37.8, built 2026-10-01 |
| docker | 28.0.4 |
| node | v22.23.3 |
| os | Linux 7.0.0-1017-azure x86_64 |
| cpus | 4 |
| memory_bytes | unknown |
| project | tdk-restaurant-example |
| healthy_expected | 8 |
| sample_vm | 0 |
Docker VM memory is only sampled on macOS runners; elsewhere it is reported as not sampled.