diff --git a/src/pages/docs/prompt/guides/run-a-prompt.mdx b/src/pages/docs/prompt/guides/run-a-prompt.mdx index 62572a257..081630863 100644 --- a/src/pages/docs/prompt/guides/run-a-prompt.mdx +++ b/src/pages/docs/prompt/guides/run-a-prompt.mdx @@ -43,7 +43,7 @@ Set the parameters that shape the model's output, things like temperature, max t ## Run the prompt and read the output -Click **Run Prompt**. The response fills in the output panel as the model generates it, rather than appearing all at once, so you can start reading before it finishes. For support-agent, a good response is a short, direct answer to `customer_question` that reads as coming from `company_name`, matching the clear, professional tone the messages ask for. +Click **Run Prompt**. If you changed the model, a parameter or a message just before, the run first waits for that change to save, so it always uses what the editor shows. The response fills in the output panel as the model generates it, rather than appearing all at once, so you can start reading before it finishes. For support-agent, a good response is a short, direct answer to `customer_question` that reads as coming from `company_name`, matching the clear, professional tone the messages ask for. If another prompt in your organization already uses the same name, the run stops before it starts. Renaming it clears the conflict. diff --git a/src/pages/docs/prompt/troubleshooting.mdx b/src/pages/docs/prompt/troubleshooting.mdx index 916428897..d67ed35d7 100644 --- a/src/pages/docs/prompt/troubleshooting.mdx +++ b/src/pages/docs/prompt/troubleshooting.mdx @@ -23,6 +23,7 @@ The blocked states and errors you actually run into in Prompt, with the cause an | "Name cannot be empty" | You tried to save a rename with the name field blank | Type a name before saving | | "You must have at least one prompt." | You tried to remove the only message left in the editor | Add a replacement message before deleting the last one | | "Audio input is missing. Please add audio before running the prompt." | You ran an audio-capable model without attaching audio content | Attach audio to the prompt, then run it | +| You press **Run Prompt** (or a compared version's **Run**) and get "The prompt changed while saving. Run it again." | While the run waited for your last change to save, you added or removed a compared version, or edited another one, so the versions no longer match what you clicked | Click **Run Prompt** again | ## Permissions, plan, and credits diff --git a/src/pages/docs/self-hosting/helm.mdx b/src/pages/docs/self-hosting/helm.mdx index 0a4a250ce..8fcb0e6c7 100644 --- a/src/pages/docs/self-hosting/helm.mdx +++ b/src/pages/docs/self-hosting/helm.mdx @@ -110,6 +110,8 @@ kubectl -n futureagi get job futureagi-bootstrap ``` Then run the same command with `helm upgrade --install` in place of `helm install`. The failed install already holds the release name, so a repeated `helm install` refuses it. + +While the job runs, the pods can already show Running and Ready: readiness and `/health/` mean the processes are up, not that the install is finished. In releases after v1.58.0, until the migrations are applied the API answers requests, including the UI's, with `503`, "The platform is still initializing; retry shortly." and a `Retry-After: 5` header, instead of failing on tables that don't exist yet. @@ -390,6 +392,7 @@ For bundled datastores, use a tested PVC backup/restore process such as Velero o | bootstrap or fi-collector: certificate verify failed | `global.caBundle`, or `PGSSLROOTCERT` and the CA mount (see [PostgreSQL](https://github.com/future-agi/future-agi/blob/main/deploy/helm/futureagi/README.md#postgresql)) | | `helm upgrade`: `... cannot change` (StatefulSet ...) | [Install-time settings](https://github.com/future-agi/future-agi/blob/main/deploy/helm/futureagi/README.md#install-time-settings) | | Workers log `waiting for the database migrations` | the bootstrap job: `kubectl -n futureagi logs job/futureagi-bootstrap` | +| API requests, including the UI's, get `503` "The platform is still initializing; retry shortly." (releases after v1.58.0) | the bootstrap job is still applying the migrations: `kubectl -n futureagi logs job/futureagi-bootstrap`. Until it does, the API logs `startup_schema_pending` at most every 5 seconds per worker. It answers normally once the migrations are applied, without a restart | | UI loads but every call fails, or CORS errors | `urls.api` or the API host; `config.corsAllowedOrigins`; with port-forwards, forward the backend to localhost:8000 | | API answers 400 Bad Request | `config.allowedHosts`: the Host header is not an allowed host | | The load balancer marks the API or collector unhealthy (`no healthy upstream`, 502, 503) | its own health checks, which skip the readiness probes: GKE `gatewayApi.gke.healthChecks`, EKS the `alb.ingress.kubernetes.io/healthcheck-*` Service annotations |