Error map
502 Bad Gateway: failed_to_connect_to_backend Google Cloud Load Balancing: 502 Bad Gateway — jsonPayload.statusDetails: failed_to_connect_to_backend / backend_connection_closed_before_data_sent_to_client
Google Cloud / Firebase · layer 6, Platform & cloud
Centre: this error. First ring: errors it is linked with (12). Outer ring: what those are linked with. Colours show the layer of a request each error belongs to. Click any dot to re-centre the map on it; hover to follow its connections. The centre dot opens the error page.
Directly related (12)
BigQuery: Access Denied: Project my-project: User does not have bigquery.jobs.create permission in project my-project
The principal can read data but is not allowed to run query jobs in the project used for billing.
Terraform: Error: Attempted to load application default credentials since neither `credentials` nor `access_token` was set in the provider block. No credentials loaded
The Google provider has no credentials (ADC, service account key, impersonation).
Terraform: Error: Error creating Network: googleapi: Error 403: Compute Engine API has not been used in project 123 before or it is disabled
Terraform called a Google API that isn't enabled in the project.
Docker push: denied: Permission "artifactregistry.repositories.uploadArtifacts" denied on resource "projects/p/locations/l/repositories/r"
Your identity cannot push or pull images in that Artifact Registry repository, or Docker isn't authenticated to the registry host.
BigQuery: Not found: Dataset my-project:analytics was not found in location EU
BigQuery looked for the dataset in the wrong location (region) or the dataset doesn't exist.
Cloud Run: 503 Service Unavailable — The request was aborted because there was no available instance
Cloud Run could not serve the request: no instance was ready, instances crashed, or the service hit its max-instances/concurrency…
GKE Ingress: Some backend services are in UNHEALTHY state / 502 Server Error from the load balancer
The Google load balancer's health checks fail for the Ingress backends, so it returns 502 or 'no healthy upstream'.
gcloud/ADC: WARNING: Your active project does not match the quota project in your local Application Default Credentials file
Application Default Credentials bill API usage to a different project than the one you have set as active, which can cause surpri…
Google Cloud: 403 SERVICE_DISABLED — Cloud Resource Manager API has not been used in project 123456 before or it is disabled. Enable it by visiting
The Google API you called is not enabled for the project that the request is billed to.
Cloud CDN not caching: cache misses because of Set-Cookie, Cache-Control or missing cache mode
Responses keep missing the CDN cache, so every request reaches the origin.
GKE Ingress stuck: Error syncing to GCP — load balancer not created / NEG not found
The Ingress controller can't create or update the Google Cloud load balancer, so the Ingress has no address or routes.
Cloud Run 502/connection error with HTTP/2 (h2c) — container doesn't speak the configured protocol
Cloud Run is configured to use HTTP/2 end-to-end, but the container only speaks HTTP/1.1 (or vice versa).
Two steps away (36)
Cloud Run: Error: Forbidden — Your client does not have permission to get URL / from this server
The Cloud Run service requires authentication and the caller has no valid identity token or the `roles/run.invoker` role.
Google Cloud: 403 PERMISSION_DENIED — Permission 'storage.objects.get' denied on resource (or it may not exist)
The caller's identity lacks the IAM permission named in the message on that resource.
GCP: FAILED_PRECONDITION: Key creation is not allowed on this service account (constraints/iam.disableServiceAccountKeyCreation)
An organisation policy forbids creating downloadable JSON keys for service accounts.
AWS: ExpiredToken — The security token included in the request is expired
Temporary AWS credentials (STS/SSO/assumed role) have expired.
AWS: SignatureDoesNotMatch / InvalidClientTokenId — the request signature we calculated does not match
AWS rejected the request signature — wrong secret key, wrong region, clock skew, or a presigned URL used differently from how it…
UnrecognizedClientException: The security token included in the request is invalid
The access key or session token isn’t recognised — wrong/deleted key, a session token missing for temporary credentials, or a reg…
Cloud Run: Memory limit of 512 MiB exceeded with 530 MiB used. Consider increasing the memory limit
The container used more memory than the Cloud Run service allows, so the instance was terminated.
Cloud Run: 504 Gateway Timeout — The request has been terminated because it has reached the maximum request timeout
The request ran longer than the service's request timeout (default 5 minutes, max 60), so Cloud Run closed it.
Cloud Run: Revision 'svc-00005' is not ready — Image 'gcr.io/proj/app:latest' not found / Permission 'artifactregistry.repositories.downloadArtifacts' denied
Cloud Run cannot pull the container image — it doesn't exist, the tag is wrong, or the service identity lacks access to the regis…
denied: requested access to the resource is denied (docker push)
You’re logged in, but not allowed to push to that repository — usually the image isn’t tagged with your own username/organisation.
gcloud: ERROR: (gcloud.auth.print-access-token) There was a problem refreshing your current auth tokens: Reauthentication failed. cannot prompt during non-interactive execution
Your cached Google credentials expired or require a fresh login (session control or password change), and gcloud can't prompt.
Google Cloud: 403 This API method requires billing to be enabled. Please enable billing on project and retry
The project has no active billing account, and the API/service needs one.
BigQuery: Resources exceeded during query execution: Not enough resources for query planning - too many subqueries or query is too complex
A query is too heavy for BigQuery to execute (usually sorts, large joins, window functions or DISTINCT over huge data on a single…
BigQuery: Syntax error / Unrecognized name: customer_id at [3:8] (Did you mean customerId?)
The query references a column or alias that doesn't exist in that scope.
BigQuery: Quota exceeded: Your project exceeded quota for free query bytes scanned. Consider enabling billing
A free-tier/sandbox project used its monthly or daily query processing allowance.
Cloud Run: The user-provided container failed to start and listen on the port defined provided by the PORT=8080 environment variable
Cloud Run started your container, but nothing listened on $PORT within the startup timeout — the app crashed, listens on another…
Cloud SQL: FATAL: remaining connection slots are reserved / MySQL: ERROR 1040 (08004): Too many connections
The instance's maximum connections was reached, often because serverless instances each open their own connection pool.
Cloud Run: Container called exit(1) — the revision failed to start
The container process exited with an error right after starting, so Cloud Run marked the revision as failed.
Readiness probe failed: HTTP probe failed with statuscode: 503 / connect: connection refused
The container is running, but its readiness check fails, so the pod is marked not ready and removed from Service endpoints.
Ingress returns "404 Not Found" / "default backend - 404" (ingress-nginx)
The ingress controller received the request but no Ingress rule matched the host/path, so it answered with its default backend.
Error from server: admission webhook "validate.nginx.ingress.kubernetes.io" denied the request
A validating admission webhook (ingress-nginx, Gatekeeper/OPA, Kyverno) inspected your object and rejected it; the reason follows…
Google Cloud: Error: Could not load the default credentials. Browse to https://cloud.google.com/docs/authentication/getting-started
The Google client library found no Application Default Credentials (ADC) on this machine or environment.
GKE/Kubernetes: Pod unschedulable — 0/3 nodes are available: 3 Insufficient cpu
The scheduler found no node with enough free CPU/memory for the pod's resource requests.
kubectl: error: executable gke-gcloud-auth-plugin not found — it looks like you are trying to use a client-go credential plugin that is not installed
kubectl needs the `gke-gcloud-auth-plugin` binary to authenticate to GKE, and it is missing from PATH.
Service has no endpoints (no endpoints available for service / Endpoints: <none>)
The Service’s selector matches no ready pods, so traffic to it is refused or returns 503.
spec.ports[0].nodePort: Invalid value: 80: provided port is not in the valid range. The range of valid ports is 30000-32767
A NodePort Service asked for a port outside the cluster’s NodePort range (default 30000–32767).
Context.startForegroundService() did not then call Service.startForeground() / ForegroundServiceStartNotAllowedException
A foreground service didn’t call startForeground() within a few seconds, or the app tried to start one from the background where…
Refused to load the script because it violates the following Content Security Policy directive
The page’s Content-Security-Policy doesn’t allow that script, style, image or connection source.
Cloudflare not caching: cf-cache-status DYNAMIC / BYPASS / MISS every time
Cloudflare isn’t caching the response — HTML isn’t cached by default, and cookies or Cache-Control headers from the origin preven…
Networking: DNS changed but the domain still resolves to the old IP — TTL and resolver caching ("propagation")
You updated a DNS record, but resolvers and devices keep serving the old answer until its TTL expires.
GKE: google.api_core.exceptions.Forbidden: 403 Caller does not have required permission — Workload Identity: unable to generate access token
A pod tried to call Google APIs but its Kubernetes service account is not bound to a Google service account (or the binding lacks…
GKE Autopilot rejected the workload: resources adjusted / violates Autopilot constraints
Autopilot modifies or rejects pods that don't meet its rules (requests/limits, privileged settings, host access).
GKE ManagedCertificate stuck Provisioning / FailedNotVisible (domain doesn't point to the load balancer)
Google can't validate the domain because DNS doesn't reach the Ingress's IP, so the certificate never becomes Active.
HTTP/2 GOAWAY received (connection closed by server)
The server sent an HTTP/2 GOAWAY frame, telling the client to stop opening new streams on this connection.
Cloud Run background work stops after the response (CPU only allocated during requests)
Threads, timers or async jobs freeze or never finish because CPU is throttled outside request handling.
Cloud Run domain mapping stuck: certificate provisioning pending (CertificatePending)
A custom domain mapped to Cloud Run never gets a certificate.
Map another error
Popular: 500400403ORA-009331064Critical errorFileNotFoundException assemblyTypeError