mockersf / concourse-slack-notifier Goto Github PK

when a job fails every time, I want to send a notification at most every X hours instead of on every failure

The notifier is not sending "fixed" information to slack anymore.
It's hard to know the cause as the notifier is not providing enough information or logs to debug the issue.

The only thing present is the resource status: "Did not send Fixed notification to #xxx"

api in resource https://github.com/concourse/docker-image-resource:

• ca_certs: Optional. An array of objects with the following format:

  ca_certs:
  - domain: example.com:443
    cert: |
      -----BEGIN CERTIFICATE-----
      ...
      -----END CERTIFICATE-----
  - domain: 10.244.6.2:443
    cert: |
      -----BEGIN CERTIFICATE-----
      ...
      -----END CERTIFICATE-----

  Each entry specifies the x509 CA certificate for the trusted docker registry
  residing at the specified domain. This is used to validate the certificate of
  the docker registry when the registry's certificate is signed by a custom
  authority (or itself).

  The domain should match the first component of repository, including the
  port. If the registry specified in repository does not use a custom cert,
  adding ca_certs will break the check script. This option is overridden by
  entries in insecure_registries with the same address or a matching CIDR.

• client_certs: Optional. An array of objects with the following format:

  client_certs:
  - domain: example.com:443
    cert: |
      -----BEGIN CERTIFICATE-----
      ...
      -----END CERTIFICATE-----
    key: |
      -----BEGIN RSA PRIVATE KEY-----
      ...
      -----END RSA PRIVATE KEY-----
  - domain: 10.244.6.2:443
    cert: |
      -----BEGIN CERTIFICATE-----
      ...
      -----END CERTIFICATE-----
    key: |
      -----BEGIN RSA PRIVATE KEY-----
      ...
      -----END RSA PRIVATE KEY-----

  Each entry specifies the x509 certificate and key to use for authenticating
  against the docker registry residing at the specified domain. The domain
  should match the first component of repository, including the port.

Hi @mockersf Thanks for this useful resource!

We noticed that the disable flag does not work properly. It returns true in should_send_message if the flag is set. I have a local fork and PR, but it seems upstream PRs are not enabled.

nevermind

Problem

Due to the planned expiration of Let's Encrypt's root certificate https://letsencrypt.org/docs/dst-root-ca-x3-expiration-september-2021/ slack formatted notifications do not work for servers with a leaf certificate provided by Let's Encrypt.

Concourse CI pipeline error

  | "error  sending request for url  (https://matrix.netresearch.de/appservice-webhooks/api/v1/matrix/hook/xxxxxx):  error trying to connect: error:14090086:SSL  routines:ssl3_get_server_certificate:certificate verify  failed:s3_clnt.c:1269: (certificate has expired)"
-- | --

Current certificate used in the Docker image

See the CN for the root certificate

/ # openssl s_client -servername matrix.netresearch.de -connect matrix.netresearch.de:443
CONNECTED(00000003)
depth=3 O = Digital Signature Trust Co., CN = DST Root CA X3
verify error:num=10:certificate has expired

Request

Is it possible for you to trigger a Docker image build and publish that to somewhere so we can continue using your great work?

In Concourse 6.0 you can now "re-run" a build. If you re-run build 314, for example, it will create a new build 314.1 consisting of the re-run. This Slack notification resource doesn't seem to like that and displays "unknown build" for these re-runs.

Screenshot:

To reproduce:

1. Use Concourse >= 6.0
2. Run a job that triggers a slack notification.
3. Re-run that job using the rerun button or the fly CLI

I'm getting the above error (relative URL without a base) when trying to set the url to something like (( slack-url )) which is stored in our credential backend.