Troubleshooting

When a request fails, Calendarpit returns an explicit status code and error message. It does not silently ignore unsupported fields or parameters.

To investigate a failure, open API history (http://127.0.0.1:4173/history) in the admin UI:

• It shows the latest 200 Calendar API and OAuth exchanges.
• Each entry records the HTTP method, path, status code, headers, and masked request and response bodies.
• The Errors toggle filters for responses with status 400 or higher.

Each entry on this page follows the same pattern: the exact Error output (when there is one), the likely Cause, and the Fix.

Error response formatsCopy Anchor Link

Google API error envelopeCopy Anchor Link

Calendar API routes (/calendar/v3/*), /userinfo, and management API routes (/__calendarpit/*) return the Google API error envelope:

{
  "error": {
    "code": 404,
    "message": "Event not found.",
    "errors": [
      {
        "domain": "calendarpit",
        "reason": "notFound",
        "message": "Event not found."
      }
    ]
  }
}

• domain is "calendarpit" on all top-level errors. Per-calendar errors inside a freebusy.query response use "domain": "global".
• 401 responses include the WWW-Authenticate: Bearer header.
• 400 validation errors without a more specific reason use "reason": "invalidParameter".

OAuth error formatCopy Anchor Link

OAuth routes (/oauth/* and /oauth2/*) return the OAuth error format. Most errors are HTTP 400, with these exceptions:

• Unimplemented paths return 501.
• Too many pending consent requests return 429.
• A cross-origin consent submission returns 403.

{
  "error": "invalid_grant",
  "error_description": "Invalid, expired, or mismatched authorization code."
}

Validation errors on GET /oauth/authorize are returned directly in the browser as HTTP 400 JSON. Calendarpit does not redirect invalid authorization requests back to redirect_uri.

Starting CalendarpitCopy Anchor Link

CLI startup failures print Calendarpit: <message> to stderr and exit with code 1. For Docker, inspect the server output with docker compose logs calendarpit.

Port already in useCopy Anchor Link

• Error: Calendarpit: Failed to start server. Is port 4173 in use?
• Cause: Another process is listening on port 4173.
• Fix: Stop that process, or choose another port with --port <n>. With --port 0, read the actual Admin URL. In Docker, change the host side of the mapping, for example 127.0.0.1:4174:4173, and open port 4174.

Database already in useCopy Anchor Link

• Error: Calendarpit: Database is already in use: <path>. Use a separate path or :memory:.
• Cause: Another process holds the lock on the same SQLite file.
• Fix: Stop that process, or give each process its own --database path or :memory:. See Storage and restarts.

Fixture file not foundCopy Anchor Link

• Error: Calendarpit: ENOENT: no such file or directory, open '<path>'
• Cause: The fixture path does not exist in the current working directory or inside the container.
• Fix: Use a valid absolute path. In Docker, mount the file and pass its container path. See Fixtures.

Fixture fails validationCopy Anchor Link

• Error: Calendarpit: Each user with calendars must have exactly one primary calendar.
• Cause: The startup fixture fails validation.
• Fix: Give each user with calendars exactly one "primary": true calendar. See the fixture rules.

Fixture edits have no effect after restartCopy Anchor Link

• Cause: Startup fixtures only seed a new database. Existing data and the reset baseline are retained.
• Fix:
  • To replace current data, load the file via Settings → Test data or POST /__calendarpit/fixtures.
  • To use a new reset baseline, start with a new database path or :memory:.

Data is missing after changing how Calendarpit startsCopy Anchor Link

• Cause: One of the following:
  • The working directory changed.
  • A different database path or volume was used.
  • The previous session used :memory:.
• Fix: Check Data <path> in startup output and Settings → Test data. Use the same absolute database path or Docker volume.

The browser, backend, or webhook receiver cannot connect in DockerCopy Anchor Link

• Cause: Loopback points to the requesting container itself. A browser cannot resolve a Compose service name.
• Fix: Use the host-published URL in the browser and the service name between containers. For webhook receivers, configure CALENDARPIT_WATCH_HOSTS. See container networking.

A CALENDARPIT_* value seems ignoredCopy Anchor Link

• Cause: A CLI flag overrides the environment variable, or the variable was set on the host but not passed to Docker.
• Fix: Remove the overriding flag, or pass the variable with docker run -e or Compose environment. When you change the container port, also update its mapping and CALENDARPIT_PORT for the health check.

Invalid option value, unknown option, or unknown commandCopy Anchor Link

• Cause: The CLI received an invalid flag, empty value, or unsupported subcommand.
• Fix: Run calendarpit --help. Only serve is supported. Ports must be integers from 0 to 65535.

OAuthCopy Anchor Link

A real Google consent screen appearsCopy Anchor Link

• Cause: Your app's authorization URL still points to Google (accounts.google.com).
• Fix: Set the authorization URL to http://127.0.0.1:4173/oauth/authorize. See Connect your app.

Token exchange or refresh contacts GoogleCopy Anchor Link

• Cause: Your OAuth client still uses Google's token endpoint. Setting the Calendar API endpoint does not redirect OAuth calls.
• Fix: Override the token URL to http://127.0.0.1:4173/oauth/token. If you use /oauth/revoke and /userinfo, override them too. See Local OAuth.

Google's default OAuth paths return 501Copy Anchor Link

• Error: 501 unsupportedFeature on /oauth2/... or /oauth/tokeninfo
• Cause: Your SDK is calling Google's default OAuth paths: /oauth2/v4/token, /oauth2/v3/userinfo, or /oauth/tokeninfo.
• Fix: Override the endpoints explicitly to /oauth/authorize, /oauth/token, /oauth/revoke, and /userinfo.

Unregistered redirect URICopy Anchor Link

• Error: invalid_request: Unregistered redirect URI.
• Cause: redirect_uri does not exactly match an entry in clients[].redirectUris.
  • The match covers localhost vs 127.0.0.1, port, path, and trailing slash.
  • Changing SAMPLE_PORT from 4175 also causes a mismatch with the built-in sample-app client.
• Fix: Copy the exact redirect_uri string into clients[].redirectUris in your fixture.

Unknown clientCopy Anchor Link

• Error: invalid_client: Unknown client.
• Cause: One of the following:
  • client_id is not registered in the fixture.
  • POST /oauth/token used a JSON body instead of application/x-www-form-urlencoded. With HTTP Basic authentication, a JSON body returns unsupported_grant_type instead.
• Fix: Register the client in your fixture. Send token requests as application/x-www-form-urlencoded.

Invalid client secretCopy Anchor Link

• Error: invalid_client: Invalid client secret.
• Cause: client_secret is missing or wrong. Calendarpit requires a client secret on every token request, including PKCE flows.
• Fix: Send the matching client_secret in the form body or via HTTP Basic authentication.

Invalid authorization codeCopy Anchor Link

• Error: invalid_grant: Invalid, expired, or mismatched authorization code.
• Cause: One of the following:
  • The code is older than 60 seconds.
  • The code was already used once.
  • redirect_uri is missing or differs from the one sent to /oauth/authorize.
  • The client differs.
  • The connection was revoked.
  • The server restarted.
• Fix: Exchange the code once within 60 seconds. Use the exact redirect_uri and client used at /oauth/authorize.

Invalid PKCE verifierCopy Anchor Link

• Error: invalid_grant: Invalid PKCE verifier.
• Cause: code_verifier does not match the S256 code_challenge, or is not 43–128 characters of [A-Za-z0-9._~-].
• Fix: Send the original verifier string used to compute the S256 challenge.

Invalid refresh tokenCopy Anchor Link

• Error: invalid_grant: Invalid refresh token.
• Cause: One of the following:
  • The connection was revoked.
  • The user was deleted.
  • The workspace was reset or a fixture was loaded.
  • An in-memory server restarted.
  • The token belongs to another client.
• Fix: Run the authorization flow again to create a new connection.

state is missing or mismatches on callbackCopy Anchor Link

• Cause: One of the following:
  • state was omitted from /oauth/authorize. It is echoed only when sent.
  • Your app started on localhost and received the callback on 127.0.0.1. Cookies holding state are not shared across hostnames.
• Fix: Pass state on /oauth/authorize. Use the same hostname (127.0.0.1 or localhost) for both your app and its redirect_uri.

Unsupported or missing scopeCopy Anchor Link

• Error: invalid_scope: Unsupported or missing scope.
• Cause: scope is empty or includes an unsupported scope, such as calendar.events.owned or Drive scopes.
• Fix: Request only supported scopes.

Scope changes on refreshCopy Anchor Link

• Error: invalid_scope: Refresh scope changes are unsupported.
• Cause: A refresh_token request included a scope parameter that does not exactly match the original grant string. Scope order counts.
• Fix: Omit scope when refreshing a token.

No refresh_token in the token responseCopy Anchor Link

• Cause: One of the following:
  • access_type=offline was not sent on /oauth/authorize.
  • The response is from a refresh_token grant. Refresh responses return a new access_token only, and the original refresh token stays valid.
• Fix: Add access_type=offline to /oauth/authorize, and keep the initial refresh_token.

Unsupported authorization parametersCopy Anchor Link

• Error: unsupported_feature on /oauth/authorize
• Cause: The request included one of the following:
  • prompt=none
  • include_granted_scopes=true
  • An unsupported parameter such as nonce, login_hint, hd, or enable_granular_consent
• Fix: Remove unsupported parameters. Use interactive consent (prompt=consent or prompt=select_account).

Invalid PKCE challengeCopy Anchor Link

• Error: invalid_request: PKCE requires S256. or Missing PKCE challenge.
• Cause: One of the following:
  • code_challenge was sent without code_challenge_method=S256 (including plain).
  • code_challenge_method was sent without code_challenge.
  • code_challenge is not 43 base64url characters.
• Fix: Use code_challenge_method=S256 with a 43-character base64url challenge.

Consent request expired or already usedCopy Anchor Link

• Error: invalid_request: Consent request expired or already used.
• Cause: One of the following:
  • The consent page was left open longer than 5 minutes.
  • The consent page was submitted twice.
  • The server restarted before submission.
• Fix: Start a fresh authorization request at /oauth/authorize.

Too many pending consent requestsCopy Anchor Link

• Error: 429 rateLimitExceeded: Too many pending consent requests.
• Cause: 1,000 consent requests are already pending. Each expires after 5 minutes. A loop in tests usually causes this.
• Fix: Complete pending consent requests or stop opening new ones, or reset the workspace.

Calendar API and SDKsCopy Anchor Link

Invalid or expired access tokenCopy Anchor Link

• Error: 401 authError: Invalid or expired access token.
• Cause: One of the following:
  • The Authorization: Bearer header is missing or malformed.
  • The access token expired (1-hour TTL).
  • The connection was revoked.
  • The workspace was reset or a fixture was loaded.
  • An in-memory server restarted.
  • A real Google token was sent.
• Fix: Refresh the token at /oauth/token. After a reset, fixture load, or revocation, authorize again.

Insufficient scopeCopy Anchor Link

• Error: 403 insufficientPermissions: Insufficient scope.
• Cause: The token lacks a required scope. For example:
  • Writing events with a .readonly scope.
  • Calling calendarList.list with only calendar.events.
  • Calling /userinfo without openid, email, or profile.
• Fix: Request a scope that covers the endpoint and re-authorize. See Local OAuth.

Unsupported API endpoint or methodCopy Anchor Link

• Error: 501 unsupportedFeature: Unsupported API endpoint or method. This includes doubled /calendar/v3/calendar/v3/... paths.
• Cause: One of the following:
  • You called an unimplemented Google endpoint: acl, settings, colors, calendars.insert, or events.move.
  • You sent HEAD to an API path. Some paths answer HEAD differently, as listed below.
  • You sent POST to /users/me/calendarList/{calendarId}.
  • The SDK base URL is misconfigured, so the SDK doubled /calendar/v3.
• Fix: Check Calendar API compatibility. Set the SDK base URL as shown in the table below. See SDK examples.

HEAD returns a different error on these paths:

• Event paths respond with 405 methodNotAllowed.
• /calendars/{calendarId} returns Unsupported calendar operation.
• An existing entry at /users/me/calendarList/{calendarId} returns Unsupported calendar list operation.

Unsupported calendar or calendar list operationCopy Anchor Link

• Error: 501 unsupportedFeature: Unsupported calendar operation. or Unsupported calendar list operation.
• Cause: One of the following:
  • You called PUT, PATCH, DELETE, or HEAD on /calendar/v3/calendars/{calendarId}.
  • You called an unsupported method such as HEAD or OPTIONS on an existing entry at /calendar/v3/users/me/calendarList/{calendarId}. POST there returns Unsupported API endpoint or method. instead.
• Fix: Manage calendars in the admin UI or via /__calendarpit/calendars. For per-user display settings, use calendarList.patch. See Test control API.

Calendar not foundCopy Anchor Link

• Error: 404 notFound: Calendar not found.
• Cause: One of the following:
  • The calendar ID is unknown.
  • The calendar is owned by a different test user.
  • You used primary for a user with no calendars.
  • The calendar was removed from the user's list with calendarList.delete. This applies to calendar-list methods only.
• Fix: Use a calendar ID owned by the authenticated user, or authorize as that calendar's owner. Restore a removed list entry with calendarList.insert.

Event not foundCopy Anchor Link

• Error: 404 notFound: Event not found.
• Cause: One of the following:
  • The event ID is unknown in that calendar.
  • You called POST .../events/import or POST .../events/quickAdd (without text). Calendarpit treats import or quickAdd as an event ID.
• Fix: Check the event ID. Use events.insert instead of import or quickAdd.

Unsupported event methodCopy Anchor Link

• Error: 405 methodNotAllowed: Unsupported event method. or Unsupported instances operation.
• Cause: The HTTP verb is invalid for the event route. For example:
  • HEAD
  • DELETE on /events
  • POST on an existing /events/{eventId}
  • A write to /instances
• Fix: Use the supported HTTP method and path.

Unsupported query parameterCopy Anchor Link

• Error: 400 unsupportedFeature: Unsupported query parameter: <key>
• Cause: The query parameter is not on the endpoint allowlist. Examples include fields, q, eventTypes, sendNotifications, supportsAttachments, maxAttendees, and text.
• Fix: Remove the parameter. Disable automatic fields partial-response masks in your SDK.

Unsupported fieldCopy Anchor Link

• Error: 400 unsupportedFeature: Unsupported field: <key>
• Cause: The request body contains an unsupported field. For example:
  • eventType, attachments, source, or htmlLink on events.
  • An items[] entry with keys other than id on FreeBusy.
• Fix: Strip unsupported fields before sending the request. See Calendar API compatibility.

Non-JSON response formatCopy Anchor Link

• Error: 400 invalidParameter: Only JSON responses are supported.
• Cause: The query included alt=media or alt=proto.
• Fix: Omit alt or use alt=json.

Timestamp without an offsetCopy Anchor Link

• Error: 400: Expected an RFC3339 timestamp with an offset. or dateTime requires an offset or timeZone.
• Cause: One of the following:
  • timeMin, timeMax, updatedMin, or FreeBusy bounds lack Z or ±HH:MM.
  • An event dateTime has neither an offset nor timeZone.
• Fix: Add Z or a ±HH:MM offset, or supply timeZone on the event start/end.

Invalid or ambiguous local timeCopy Anchor Link

• Error: 400: Invalid or ambiguous local time; supply an explicit offset.
• Cause: An offset-less dateTime falls inside a DST gap or overlap in timeZone.
• Fix: Include an explicit numeric offset on dateTime.

Recurring event time zone or UNTIL errorsCopy Anchor Link

• Error: 400: Recurring timed events require start.timeZone. or Timed recurrence UNTIL must be UTC.
• Cause: A timed recurring event has no start.timeZone, or its UNTIL value does not end with Z.
• Fix: Set start.timeZone. Write timed UNTIL timestamps in UTC (...Z). See Events, recurrence and time.

Unsupported recurrence rule or expansionCopy Anchor Link

• Error: 400 unsupportedFeature on recurrence rules or expansion
• Cause: One of the following:
  • The rule uses HOURLY or MINUTELY.
  • The rule uses an unsupported part: BYHOUR, BYWEEKNO, or BYYEARDAY.
  • The rule includes another iCalendar line (DTSTART).
  • An occurrence falls in a DST gap.
  • An unbounded rule is expanded without timeMax.
  • The expansion exceeds 10,000 instances or 100 years.
• Fix: Use supported DAILY, WEEKLY, MONTHLY, or YEARLY rules. Supply timeMax and a narrower window. See Events, recurrence and time.

orderBy=startTime without singleEventsCopy Anchor Link

• Error: 400: startTime ordering requires singleEvents=true.
• Cause: orderBy=startTime was sent without singleEvents=true.
• Fix: Add singleEvents=true to the events.list query.

Event is not a recurring seriesCopy Anchor Link

• Error: 400: Event is not a recurring series.
• Cause: events.instances was called on a non-recurring event.
• Fix: Call events.instances only on recurring series masters.

Read-only field changedCopy Anchor Link

• Error: 400: <key> is read-only.
• Cause: One of the following:
  • A Calendar API write changed organizer, creator, iCalUID, recurringEventId, or originalStartTime.
  • A calendarList write changed summary, timeZone, accessRole, primary, kind, deleted, or conferenceProperties.
• Fix: Omit the field or send the existing value unchanged.
  • For a per-user calendar name, set summaryOverride.
  • To set custom organizer metadata, use Fixtures or the Test control API.

conferenceData without conferenceDataVersionCopy Anchor Link

• Error: 400: conferenceData requires conferenceDataVersion=1.
• Cause: The event payload includes conferenceData without ?conferenceDataVersion=1.
• Fix: Add conferenceDataVersion=1 to the query string.

Cannot remove the primary calendarCopy Anchor Link

• Error: 403 cannotDeletePrimaryCalendar: Cannot remove the primary calendar from the calendar list.
• Cause: calendarList.delete was called on the user's primary calendar.
• Fix: Keep the primary calendar in the list. To hide it, use calendarList.patch with hidden: true or selected: false.

Invitee edits are rejectedCopy Anchor Link

• Error: 403 forbiddenForNonOrganizer
• Cause: An invitee editing an invitation copy did one of the following:
  • Used PUT.
  • Changed shared event fields.
  • Edited another attendee.
• Fix: Use PATCH. Update only your own responseStatus/comment, reminders, colorId, transparency, or extendedProperties.

Duplicate event or calendar list entryCopy Anchor Link

• Error: 409 duplicate with one of these messages:
  • Duplicate event ID.
  • Invitation event ID conflicts with an existing local event.
  • Calendar already exists in the calendar list.
• Cause: One of the following:
  • events.insert reused an existing ID in that calendar.
  • An attendee invitation copy matched an event ID in the attendee's primary calendar with a different iCalUID. For example, the existing event's iCalUID was set through a fixture or the Test control API. An existing event with the default iCalUID is overwritten instead.
  • calendarList.insert was called for a calendar already in the list, including primary.
• Fix: Use a unique event ID. Call calendarList.insert only for a calendar that is not in the list.

Event is already deletedCopy Anchor Link

• Error: 410 deleted: Event is already deleted.
• Cause: events.delete was called on an event that is already cancelled.
• Fix: Treat 410 on delete as already deleted.

etag mismatchCopy Anchor Link

• Error: 412 conditionNotMet: Event has changed. or Calendar list entry has changed.
• Cause: The If-Match header does not match the current etag of the event or calendar-list entry.
• Fix: Re-fetch with events.get or calendarList.get and retry.

Request body too largeCopy Anchor Link

• Error: 413 payloadTooLarge: Request body too large.
• Cause: The request body exceeds 1,048,576 bytes (1 MiB).
• Fix: Reduce the payload size.

maxResults out of rangeCopy Anchor Link

• Error: 400: maxResults must be between 1 and 2500.
• Cause: maxResults is 0, negative, or greater than 2500.
• Fix: Set maxResults between 1 and 2500.

Python SDK fetches discovery documents from GoogleCopy Anchor Link

• Cause: googleapiclient.discovery.build tried to download discovery metadata from Google.
• Fix: Use google-api-python-client 2.x with static_discovery=True and cache_discovery=False. Do not set a remote discovery URL. See SDK examples.

Incremental sync and paginationCopy Anchor Link

Sync token is invalid or expiredCopy Anchor Link

• Error: 410 fullSyncRequired: Sync token is invalid or expired. Perform a full synchronization.
• Cause: One of the following:
  • The syncToken is older than 7 days.
  • The syncToken was evicted because the workspace has more than 100 tokens.
  • The syncToken belongs to another calendar or user.
  • A reset, fixture load, or in-memory restart cleared the syncToken.
• Fix: Discard local sync state and run a full events.list without syncToken. See Incremental sync.

Sync query options changedCopy Anchor Link

• Error: 400: Sync query options must match the initial synchronization.
• Cause: singleEvents or the effective timeZone differs from the initial sync. When timeZone was omitted, a calendar time zone change also counts.
• Fix: Repeat the exact singleEvents and timeZone options from the initial full sync.

Invalid syncToken combinationCopy Anchor Link

• Error: 400 with one of these messages:
  • <key> cannot be combined with syncToken.
  • Incremental synchronization must include deleted events.
  • instances does not support syncToken.
• Cause: One of the following:
  • syncToken was combined with timeMin, timeMax, orderBy, updatedMin, iCalUID, or extended property filters.
  • syncToken was sent with showDeleted=false.
  • syncToken was used on events.instances.
• Fix: Send only syncToken on events.list. You may also send matching singleEvents, timeZone, maxResults, pageToken, or showDeleted=true.

Invalid or expired page tokenCopy Anchor Link

• Error: 410 invalidPageToken: Invalid or expired page token. or Invalid or stale page token.
• Cause: The page token is no longer valid.
  • An event pageToken expired (1-hour TTL), was evicted (over 1,000 frozen pages), or was invalidated because query options changed or the server restarted or reset.
  • A calendarList pageToken was used after any workspace data write.
• Fix: Restart listing from the first page.

Watch notificationsCopy Anchor Link

Watch host is not allowedCopy Anchor Link

• Error: 400: Watch host is not allowed. Configure CALENDARPIT_WATCH_HOSTS explicitly.
• Cause: The callback address hostname is not 127.0.0.1, [::1], or localhost. For example, it is host.docker.internal or a container service name.
• Fix: Start Calendarpit with CALENDARPIT_WATCH_HOSTS=<host1>,<host2>. List hostnames only, without ports. See Watch notifications.

Invalid watch callback URLCopy Anchor Link

• Error: 400: Invalid watch callback URL.
• Cause: address is not http: or https:, or includes credentials or a #fragment.
• Fix: Use a plain http:// or https:// callback URL.

Invalid channel fieldsCopy Anchor Link

• Error: 400 with one of these messages:
  • Only web_hook channels are supported.
  • Invalid channel ID.
  • Invalid channel token.
  • Invalid expiration.
• Cause: One of the following:
  • type is not "web_hook".
  • id does not match [A-Za-z0-9_-]{1,64}.
  • token exceeds 256 characters or contains newlines.
  • expiration is not a future Unix millisecond integer.
• Fix: Fix the invalid field in the events.watch or calendarList.watch request body.

Duplicate channel IDCopy Anchor Link

• Error: 409 duplicate: Duplicate channel ID.
• Cause: Another active watch channel in the workspace already uses that id.
• Fix: Generate a new unique id (such as a UUID) for each channel.

Watch succeeds, but no notifications arrive at your appCopy Anchor Link

• Cause: Delivery to your webhook receiver failed. After 3 attempts, the delivery is marked failed. An attempt fails when the receiver:
  • Is unreachable.
  • Takes more than 2,000 ms.
  • Returns a status other than 200, 201, 202, or 204. 3xx redirects are not followed.
• Fix: Open Watch (http://127.0.0.1:4173/watch) in the admin UI to inspect the delivery log status and HTTP status code. In tests, call POST /__calendarpit/deliveries/drain to wait for pending deliveries.

Notifications stop arriving after restart, reset, or revocationCopy Anchor Link

• Cause: One of the following:
  • Restarting cancels queued deliveries.
  • An in-memory restart drops channels.
  • A restart with a disallowed callback host drops channels.
  • Resetting, loading a fixture, or revoking the connection stops channels.
  • Deleting a calendar stops the event channels on that calendar.
  • The channel expired. Channels expire after 24 hours by default, and after 7 days at most.
• Fix: Re-fetch state after a restart. If the channel was stopped or expired, register a new watch channel.

Channel not found on channels.stopCopy Anchor Link

• Error: 404 notFound: Channel not found. on channels.stop
• Cause: One of the following:
  • The resourceId is wrong.
  • The channel ID is unknown.
  • The channel was already stopped, or pruned after expiry.
  • channels.stop was called with a token from a different connection than the one that created the channel.
• Fix: Call channels.stop with the creating connection's token and the matching resourceId. An expired channel that has not yet been pruned still returns 204. See Watch notifications.

Management APICopy Anchor Link

Management writes require JSONCopy Anchor Link

• Error: 415 unsupportedMediaType: Management writes require application/json.
• Cause: A POST, PATCH, PUT, or DELETE request to /__calendarpit/* was sent without Content-Type: application/json. This includes bodyless calls like /reset or /deliveries/drain.
• Fix: Include -H 'Content-Type: application/json' on every management write request. See Test control API.

Cross-origin management requestCopy Anchor Link

• Error: 403 forbidden: Cross-origin management requests are not allowed.
• Cause: A browser script on another origin called /__calendarpit/*.
• Fix: Call /__calendarpit/* from your test runner, CLI, or backend, without a cross-origin Origin header. Or use the admin UI.

State response posted as a fixtureCopy Anchor Link

• Error: 400 unsupportedFeature: Unsupported field: storage on POST /__calendarpit/fixtures
• Cause: You posted a GET /__calendarpit/state response directly as a fixture.
• Fix: Post a valid fixture object containing only users, calendars, events, and clients. Include the client secret values. See Fixtures.