Shared AWS HTTP API Gateway


I have a REST API Gateway set up, shared across a number of services, and am attempting to move it to an HTTP Gateway. The gateway is defined in one Serverless stack and exports a value which is imported into other stacks.

With the REST version, I get the stages and integrations set up automatically by Serverless, but not so with the HTTP GW, resulting in ‘{“message”:“not found”}’ responses to every call.

I can go into the AWS console and manually create a $default stage and point it at the Lambdas, but this is a very big change from the earlier behaviour, and not something I want to be doing.

In fairness, the Serverless documentation for reusing HTTP APIs in different services does say that these APIs, stages, and resources will not be created when importing values.

My question is “why”? Am I missing some fundamental point of HTTP APIs which now makes importing these values and creating these stages totally unnecessary? Am I wrong to expect that stage-creation behaviour to be the same as with APIv1 gateways?

I should add that when I take the ID for the HTTP API and hard-code it into the services, everything works as expected.

Shameless bump

I know it is Christmas time and all, but is there someone from the Serverless Framework project who might be able to answer this? I’d love to move to httpApis, but can’t yet as I can’t use my shared gateway.

There doesn’t seem to be a whole lot of activity in here, sadly, and unfortunately I don’t know the answer to your question or I’d try to help.
You could try over at the serverless gitter -

I’ve asked on Twitter whether this is still supposed to be the best place to come for help -

Another place you could try is the serverless channel at

That seems to get more frequent replies