PR #134 works around an Azure Container Apps limit: a probe's InitialDelaySeconds is capped at 60, and Defang maps compose's start_period onto it. compose.yaml's app healthcheck was start_period: 240s, so the deploy failed outright. Fix in that PR: start_period: 60s / retries: 10.
pulumi-defang PR 536 (merged 2026-09-05) fixes the clamp at the provider level, so this compose.yaml workaround becomes unnecessary once a release past that merge ships. As of now the latest release is v2.6.0 (2026-08-16), before 536 — releases are blocked on #414 (auto-tag workflow, needs RELEASE_PAT).
Action once a release including PR 536 ships: restore compose.yaml's app healthcheck to retries: 5 / start_period: 240s (or whatever value fits) and drop the explanatory comment.
Filed per review discussion on #134: #134 (comment)
PR #134 works around an Azure Container Apps limit: a probe's
InitialDelaySecondsis capped at 60, and Defang maps compose'sstart_periodonto it.compose.yaml's app healthcheck wasstart_period: 240s, so the deploy failed outright. Fix in that PR:start_period: 60s/retries: 10.pulumi-defangPR 536 (merged 2026-09-05) fixes the clamp at the provider level, so this compose.yaml workaround becomes unnecessary once a release past that merge ships. As of now the latest release is v2.6.0 (2026-08-16), before 536 — releases are blocked on #414 (auto-tag workflow, needsRELEASE_PAT).Action once a release including PR 536 ships: restore
compose.yaml's app healthcheck toretries: 5/start_period: 240s(or whatever value fits) and drop the explanatory comment.Filed per review discussion on #134: #134 (comment)