Is your feature request related to a problem? Please describe. I don't use Slack or Gotify.

Describe the solution you'd like I'd like to be able to specify an arbitrary set of Telegram bot API credentials, as well as an arbitrary chat ID. Messages should arrive at that chat.

Describe alternatives you've considered I tried looking for a Tekegram plugin for Gotify, but was unsuccessful.

If you are open to a suggestion, I would probably make a few tweaks to the config.yml structure to make it a little easier to navigate for new users. I realize this would be a breaking change. Maybe it could be incorporated into the version that adds settings thru the GUI. First step is to validate the config.xml structure and adjust accordingly. Below would be my suggestion.

Current Example

  Sonarr/Sonarr:
    type: url
    url: http://services.sonarr.tv/v1/releases
    semantic_versioning: false
    url_commands:
      - type: regex
        regex: \"version\":\s?\"([0-9.]+)\"
        index: 1
    web_url: http://osu.sonarr:8079/sonarr/system/updates
    icon: https://raw.githubusercontent.com/NX211/homer-icons/master/png/sonarr.png
    deployed_version:
      url: http://osu.sonarr:8079/sonarr/api/v3/system/status?apikey=428f9a8d1420426ca21a891c928594b5
      json: version

Proposed Example

  Sonarr/Sonarr:
    options:
      semantic_versioning: false
    latest_version:
      type: url
      url: http://services.sonarr.tv/v1/releases
      url_commands:
        - type: regex
          regex: \"version\":\s?\"([0-9.]+)\"
          index: 1
    dashboard:
      web_url: http://local.sonarr:8079/sonarr/system/updates
      icon: https://raw.githubusercontent.com/walkxhub/dashboard-icons/master/png/sabnzbd.png
    deployed_version:
      url: http://local.sonarr:8079/sonarr/api/v3/system/status?apikey=<redacted>
      json: version

Please consider support for local webhooks such as webhook tool. It would also be extra-useful to somehow configure argus to change an HTTP Query parameter so different services can call the same webhook with different parameters. Such as for instance: http://192.168.9.9/hooks/update-docker?container=argus.

@JosephKav I was going to wait until Shoutrr integration is implemented, but thought I would give Gotify a spin. I am not getting notifications. Am I missing something?

Global Block

settings:
  web:
    listen_host: 0.0.0.0
    listen_port: "8111"
    route_prefix: /
defaults:
  service:
    access_token: <redacted>
    interval: 30m
    deployed_version: {}
gotify:
  GOTIFY_NOTIFICATION:
    url: http://osu.gotify:8132
    token: <redacted>
    title: Argus Notification
    message: {{ service_id }} - {{ version }} released.
    extras:
      client_display: text/plain
      client_notification: '{{ web_url }}'
    priority: 5
    delay: 0s
    max_tries: 3

Service Definition

Service:
  CaddyDockerProxy:
    type: url
    url: https://github.com/lucaslorentz/caddy-docker-proxy/tags
    semantic_versioning: false
    url_commands:
      - type: regex
        regex: releases\/tag\/(v[0-9.]+)
    web_url: https://github.com/lucaslorentz/caddy-docker-proxy/releases
    icon: https://raw.githubusercontent.com/walkxhub/dashboard-icons/master/svg/caddy.svg
    gotify:
      GOTIFY_NOTIFICATION: {}
    deployed_version:
      url: http://osu.logs:8069/caddy2/current_version.txt
      regex: caddy-docker-proxy/plugin/v2\s?(v[0-9.]+)

webhooks currently fail to send if the cert is invalid (e.g. self-signed). This'll add an allow_invalid_certs key to each webhook (with support for it in defaults/globally like notify)

Triggered by: Originally posted by @samcro1967 in https://github.com/release-argus/Argus/issues/61#issuecomment-1139014316

So for fun (I mean who doesn't love certs), I changed webhook to use https. I get a different error in the webhook log when I execute the webhook from Argus, but I am guessing it is related to how I generated the self signed cert. Thought I would share in case it sheds any new light on the topic.

webhook log

[webhook] 2022/05/26 15:30:37 http: TLS handshake error from 172.18.0.14:41156: tls: oversized record received with length 21536

cert generation command

openssl req -new -x509 -sha256 -newkey rsa:2048 -nodes -keyout key.pem -days 9999 -out cert.pem`

I am getting an error using a positive look behind. It works at https://regex101.com/. I tried various escapes for the ? and <. Any ideas?

Logs

panic: regexp: Compile(`(?<=\"version\":\s\")([0-9.]+)\"`): error parsing regexp: invalid or unsupported Perl syntax: `(?<`

Regex

(?<=\"version\":\s\")([0-9.]+)\"

If I understand the current webhook implementation, once a new version is detected and a webhook is successfully sent, you can not resend it. Restarting Argus reenables the ability to send it again. I can resend a webhook on a service whose deployed version matches the current version. This seems backward to me. I want to resend a webhook on a service that is still not up to date, not on a service that is already up to date. Maybe I am thinking about this in the wrong way???

I have services setup to auto approve with a delay of one hour. This seems to work 90+ percent of the time. Occasionally even though the webhook is successfully sent and executed, the version is still not updated. This is happening as the updates have not replicated throughout CDN when the webhook is pulling the latest docker image as an example. This time will vary obviously so I settled on an hour as it works most of the time.

Ideally I think from a process flow it would look like this when auto approve is enabled:

Interval check
New version is detected
Delay times out
Webhook is executed
Interval check
Version is still not up to date
Webhook is executed again
Interval check
Version is still not up to date
Webhook is executed again
Interval check
Version is up to date

Would probably want some max retry around the webhook upgrade executions different than retrying the webhook because it failed to send the webhook. After a few failures, human intervention will probably be required. No point in staying in that hamster wheel forever. Thoughts?

I'd like to request a feature where in instead of configuring all of the services as part of the primary config file, that the primary config file pull them from any .yml file in a given services folder.

I've been playing around with services all afternoon and got 3 working, after which I commented them all out in the config file in order to work on a 4th (logging was getting very busy) only to realize that the server overwrite / deleted all commented lines in the config file thus resulting in me needing to start from scratch on the 3 I had already got working.

Having each service be it's own file would allow for easy enabling/disabling by renaming the files to a non .yml extension and would also just keep everything a lot more organized when you end up with dozens of monitored services (my goal is to use this to monitor all server based applications we host eventually which could end up being quite a lot)

Hi there,

I am very interested in this project and have been trying to set up a test vm with it for awhile but i've been struggling to get it up and running with the getting started documentation and was hoping to get some assistance/ request (or even help create) better setup documentation once I understand how to get going:

what i've done so far:

downloaded latest binary Argus-0.6.0.linux-amd64

transferred binary to centos 8 vm via scp

gave the binary execute permissions via chmod

ran ./Argus-0.6.0.linux-amd64 -help

created a /home/argus/argus directory structure and moved the binary to /home/argus/argus/Argus-0.6.0.linux-amd64

created a config.yml at /home/argus/argus/config.yml

added the following to the config.yml file just for a test:

settings:
  log:
    level: INFO           # Log level, DEBUG/VERBOSE/INFO/WARNING/ERROR
    timestamps: false     # Log with timestamps
  web:
    listen_host: 172.29.2.128  # IP address to listen on
    listen_port: 8080     # Port to listen on
    route_prefix: /       # Web route prefix, e.g. /demo means http://IP:PORT/demo to access
    cert_file: ''         # HTTPS Cert path, e.g. `cert.pem`
    pkey_file: ''         # HTTPS PrivKey path, e.g. `privkey.pem`

service:
  adnanh/webhook:
    type: github
    url: adnanh/webhook
    web_url: https://github.com/adnanh/webhook/releases/{{ version }}
    regex_content: webhook-linux-amd64\.tar\.gz
    icon: https://raw.githubusercontent.com/adnanh/webhook/development/docs/logo/logo-128x128.png

created a systemd service file at /etc/systemd/system/argus.service

[Unit]
Description=Argus Server
Documentation=https://release-argus.io/docs
After=network-online.target

[Service]
User=argus
Restart=on-failure
ExecStart=/home/argus/argus \
  -config.file=/home/argus/argus/config.yml

[Install]
WantedBy=multi-user.target

I'm not sure what else, but i feel like i'm definitely missing something.... i cant hit the web ui at the boxes ip address (172.29.2.128:8080) also tried disabling firewall and se linux so it's neither of those

Hello,

I am trying to use the new shoutrrr notification to send telegram messages. I have created an entry in:

defaults:
  notify:
    telegram_updates:
      type: telegram
      url_fields:
        Token: TOKEN_FROM_BOTFATHER
      params:
        Chats: 'chat_id'

• Do I need now to add a "notify:" entry for each service for them to notify on new releases? I haven't and I don't receive any notifications. (makes sense as I only want to receive notifications on certain services not all).
• If I add an entry "notify: telegram_updates:" in a service it complains that the type is needed, if I then add the type it asks for token, so it seems I need to repeat the config in defaults.

Is this the expected behaviour? Regards,

It appears that each time Argus is restarted the Github APIs are queried. Restarting Argus several times when working on adding a new service runs into Github rate limiting pretty easily. This maybe become a moot point if #4 is implemented so service can be added dynamically thru the UI.

Maybe rather than just querying the Github API at Argus start, check the timestamp of the last check, the interval in config.yml if defined, and delay checking until it is time to do so to avoid rate limiting.

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

Bumps @commitlint/cli from 17.3.0 to 17.4.1.

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

Bumps json5 from 1.0.1 to 1.0.2.

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

Bumps @commitlint/config-conventional from 17.3.0 to 17.4.0.

Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting @dependabot rebase.

Is your feature request related to a problem? Please describe. I am planning to build a Helm Chart for the deployment of Argus.

I would make it available, as a subfolder called chart or helm.

Looking for a way to get a list of the services that are currently being monitored by Argus from the cli. I know they are printed in the logs when the container starts, but hoping there is another way to grab them without having to parse docker logs. :-)

I can run docker exec -it argus bash -c 'argus' > services.txt and get them that way. It tries to start a new instance of argus and fails as the port is already in use, but it does print out the services that are active in config.yaml. Probably not the best way to go about it.

Would it be possible to add another cli option such as -config.services that simply lists all of the monitored services each on its own line similarly to how it does in the logs?