* Add new veracode scan for schedueled workflow

* Re-add - check_outdated job

#24)

…ks (#27)

* Updating dependencies and adding slack notifications to outdated checks

* Add slack orb

…n established in prison staff hub (#28)

* Ensure that the `SLACK_ACCESS_TOKEN` env var is set for `check_outdated`

As the final step (on fail) is a slack notification we need to use a context containing the variable.

* Refactor the channel for alerts into parameters

This establishes a reusable pattern in case consumers wish to notify slack on other jobs.

The "medium" configuration was not an acceptable term, should be one of
"info", "low", "moderate", "high" or "critical". This meant that it
would not pick up on anything!

* Updating dependencies and node

* Use parameterised slack channel rather than default of typescript alerts channel

* Extracting out executor

* WFP-322 use the hmpps/node executor to build

* WFP-322 parameterise the node version in the executor

* WFP-322 use node version parameter in integration test image

* WFP-322 integration tests use new node_redis executor

Co-authored-by: Andrew Lee <1517745+andrewrlee@users.noreply.github.com>

* Deploy template project to template k8s namespace

* DT-2404: 💄 Switch to sign in / out instead of login / out

* DT-2404: 💄 Don't need .gitignore for husky any more

…ci), and weekly policy scan (upload to veracode portal). (#45)

…imiting at veracode. (#47)

* Remove use of semi-colons before arrays

* Fix typos in README

* WFP-610 update to npm 7

* WFP-610 update outdated dependencies

* WFP-610 fixed some audit vulnerabilities

* WFP-610 updated passport-oauth2

* Moving to use HMPPS header

* Removing explicit reference to DPS

and properly indent "build" job (jobs should be an array of [name] to dictionary)

Co-authored-by: Jon Brighton <jonathan.brighton@digital.justice.gov.uk>

…ng the user directly and not from the request because passport isn't in the test stack (#62)

Also fixing open handle in test and bumping dependencies

#70)

* ⬆️ update dependencies

* 💄Add no-only-tests linting rule for cypress

Need to refer to build args before calling apt-get upgrade otherwise the set of packages are cached and not upgraded. Docker cannot cache anything in layers after a dynamic variable has been used

Also bumping version of node and fixing test compilation issue

This change sets up prometheus metrics to be available on port 3001, and
with the helm chart changes they will automatically get scraped and be
available for alerts and dashboards in grafana.

The added metrics include:

- General nodejs stats: memory use, gc etc
- HTTP server requests: counters and timings of all served HTTP requests
  by the app.
- HTTP client requests: counters and timings of all HTTP requests to
  other upstream APIs (as long as they are based off `restClient.ts`).
- Upstream healthchecks: guages recording the status/health of each
  upstream service when the healthcheck is tested.

This is all backported from the `manage-recalls-ui` app, please let me
know what you think. :)

Bumps [minimist](https://github.com/substack/minimist) from 1.2.5 to 1.2.6.
- [Release notes](https://github.com/substack/minimist/releases)
- [Commits](https://github.com/substack/minimist/compare/1.2.5...1.2.6)

---
updated-dependencies:
- dependency-name: minimist
  dependency-type: indirect
...

Signed-off-by: dependabot[bot] <support@github.com>

Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

Co-authored-by: sp-watson <steve.watson@digital,justice.gov.uk>

… out (#81)

Co-authored-by: Jon Wyatt <>
Co-authored-by: Andrew Lee <1517745+andrewrlee@users.noreply.github.com>

* SDI-181: 🔧 Use new cimg redis executor

* SDI-181: 🔧 Improve docker ignore and use released orb instead

* SDI-181: ⬆️ Bump node minor version

… tests (#86)

This caused an issue with our build where we received a segmentation fault as soon as the integration tests called the node app. Segmentation faults generally indicate an issue with one of the native C/C++ modules and it appears that one of these modules was relying on something we have stashed in the cache.

* SDI-88: 🚨 Fix querystring warning

* SDI-88: 🚨 Second attempt to querystring warning

…et method (#90)

* INC-567: Remove unnecessary type assertions

* INC-567: Return resolved promise directly

…rt major upgrade (#96)

* Ignore false positive around nodemon

* Manage version of audit ci and provide full path to ignored dependency

This allows passing through a container of wired up services through to route

This approach has been used for the dps-shared and farsight projects and it leads to a testing approach that scales more naturally. Means you can pass through the services through to where they are needed and this can grow without changes propagating through the application

Also extracted standard router into standalone middleware as the current approach relies on mutation and encourages making multiple copies of it.

The govuk/template.njk which the layout.njk extends which these files use already includes a <main> element

According to the HTML spec there should only be one <main> element present in the document at a time

* Bumping dependencies and fixing page width

There seems to be a lot of variability in page width so going with something that seems most popular in HMPPS

* Run tests in band

Partially to fix tests hanging in circle, but also as test seems to run almost twice as fast (after clearing cache)

Based on what I’ve seen elsewhere, this seems to now be a common
approach to allow us to inline scripts, see:

https://content-security-policy.com/nonce

The GOV.UK frontend has now been updated to support the use of the
`cspNonce` local - see:

 alphagov/govuk-frontend@2e40d74

So we no longer need to add the inline hash, which is vulnerable to if
the code in the frontend template is changed.

I’ve also removed the domain-specific overrides for jQuery scripts and
styles, as we can use the nonce for this too.

This speeds up the running of jest tests by enabling isolatedModules
which has the effect of [disabling typechecking](https://kulshekhar.github.io/ts-jest/docs/getting-started/options/isolatedModules)

It also drastically reduces memory usage, allowing for running tests in
parallel locally at least.

On my laptop this reduces the time to run the tests in this project from
~14 seconds to ~4 seconds. On larger projects the effect is much more pronounced,
welcome-people-to-prison reduces build time from ~2mins, 20 seconds to ~25
seconds. In circle we still need to run in band but this is still
significantly faster than before, in WPIP it reduces the build by over 1 min.

Type checking is still available in the IDE, it is also part of the
husky pre-commit hook and run by circle as part of the build, so the
risk of type errors slipping through are very small.

(We could possibly add a typechecking stage before running jest and it
would still be much faster but not adding unless it becomes apparent
that we need it )

It would be worth to re-assess this after jest 29 as there seems to be
some fixes around a [memory issue](jestjs/jest#11956) that is part of node in versions
> 16.10

NB: This will not work if type declaration (`d.ts`) files contain enums
or any other constructs that generate javascript code. This seems to be
a bit of an anti-pattern anyway (see [here](https://lukasbehal.com/2017-05-22-enums-in-declaration-files/)).
Other tooling such as cypress will only allow you to import
types from these files.