Shared code for central services
The shared sendRequest helper (src/util/request.js) configures the default outbound http.Agent with a bounded connection pool to prevent outbound connection churn under sustained high load. The pool is configurable via the following environment variables:
| Variable | Default | Description |
|---|---|---|
HTTP_AGENT_KEEP_ALIVE |
true |
Whether to keep sockets alive for reuse. Set to false to disable. |
HTTP_AGENT_MAX_SOCKETS |
512 |
Maximum number of sockets per host. Bounds the pool so bursty concurrency queues requests instead of opening unbounded sockets (node's default is Infinity). |
HTTP_AGENT_MAX_FREE_SOCKETS |
512 |
Maximum number of idle (free) sockets kept warm per host (node's default is 256). |
HTTP_AGENT_KEEP_ALIVE_MSECS |
30000 |
Initial delay (ms) for TCP Keep-Alive packets on kept-alive sockets. |
Numeric values are parsed with Number(...) and fall back to the defaults above when unset or non-numeric.
This repository uses the mojaloop/build CircleCI orb for standardized CI/CD workflows, including automated Grype vulnerability scanning for source code security.
We use audit-ci along with npm audit to check dependencies for node vulnerabilities, and keep track of resolved dependencies with an audit-ci.jsonc file.
To start a new resolution process, run:
npm run audit:fixYou can then check to see if the CI will pass based on the current dependencies with:
npm run audit:checkThe audit-ci.jsonc contains any audit-exceptions that cannot be fixed to ensure that CircleCI will build correctly.
As part of our CI/CD process, we use a combination of CircleCI, standard-version npm package and github-release CircleCI orb to automatically trigger our releases and image builds. This process essentially mimics a manual tag and release.
On a merge to main, CircleCI is configured to use the mojaloopci github account to push the latest generated CHANGELOG and package version number.
Once those changes are pushed, CircleCI will pull the updated main, tag and push a release triggering another subsequent build that also publishes a docker image.
-
There is a case where the merge to main workflow will resolve successfully, triggering a release. Then that tagged release workflow subsequently failing due to the image scan, audit check, vulnerability check or other "live" checks.
This will leave main without an associated published build. Fixes that require a new merge will essentially cause a skip in version number or require a clean up of the main branch to the commit before the CHANGELOG and bump.
This may be resolved by relying solely on the previous checks of the merge to main workflow to assume that our tagged release is of sound quality. We are still mulling over this solution since catching bugs/vulnerabilities/etc earlier is a boon.
-
It is unknown if a race condition might occur with multiple merges with main in quick succession, but this is a suspected edge case.