Hi folks - a few small, reproducible issues I ran into this week while integrating with the Verda Cloud API and this forum. Happy to provide raw request/response captures for anything below.
1. Public OpenAPI spec documents an endpoint that now returns 410
GET /v1/instance-types/price-history is still documented in the public spec (Verda Cloud Public API → /v1/openapi.json), but the live API returns 410 {"code":"this_endpoint_has_been_removed_dynamic_pricing_is_no_longer_supported"}. Anyone doing codegen/SDK work straight from the spec will hit a dead endpoint - worth dropping it from the spec (or flagging the removal in the schema).
2. Misleading validation message for well-formed UUIDs that simply don’t exist
GET /v1/instances/00000000-0000-0000-0000-000000000000 → 400 {"message":"instanceId must be a UUID"}. The id is a UUID - a 404 (or a message like “instance not found”) would save integrators a debugging session.
3. “Log in with Discourse ID” on this forum errors out (redirect_uri mismatch)
Clicking Log in with Discourse ID on /login or /signup lands on id.discourse.com and returns:
An error has occurred - The requested redirect uri is malformed or doesn’t match client redirect URI.
Reproduced several times on both intents. Looks like the OAuth client’s allowed redirect URIs need updating. (Local signup and the email login-code path still work, so it’s not a full blocker - but it’s likely the first thing new users click.)
4. Community mail quirk: the activation email never arrived
My account sat unactivated because the activation email never showed up (checked inbox + spam). The login-code email arrived instantly - so outbound mail works; might be worth checking the activation template/queue.
5. Minor: TLS SAN mismatch on cmk.datacrunch.io
curl https://cmk.datacrunch.io/rest/settings fails with no alternative certificate subject name matches 'cmk.datacrunch.io' - the certificate doesn’t include the hostname it’s served on (works with -k). Monitoring tooling can’t verify it as-is.
Separately: I recently wrote a public comparison of Verda’s GPU pricing/cost-efficiency for ML workloads (also submitted through the content program - gpuecon.substack.com). If there is a feedback/credits program for this kind of contribution, I’d be glad to keep producing technical write-ups like this - happy to be pointed in the right direction.
Thanks!