Enables Phoenix to send PING jobs to agents and receive monitoring
check results, fixing the "unknown" status for devices assigned to
local (non-cloud) agents.
Changes:
- Rename JobType::MONITOR to JobType::PING in protobuf for clarity
- Add validate_monitoring_check_message() to Validator module
- Add "monitoring_check" channel handler to receive ping results
- Implement store_monitoring_check() to save results to database
- Add build_ping_job() to generate PING jobs for agents
- Update build_jobs_for_device() to include ping jobs when
monitoring_enabled=true
Phoenix now sends PING jobs alongside SNMP/MikroTik jobs, and agents
send back MonitoringCheck results that are stored in the
monitoring_checks table with proper agent_token_id tracking.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
When an organization has sites disabled (use_sites=false), the API
now automatically removes any site_id from device creation requests,
ensuring devices are assigned directly to the organization instead.
This prevents Terraform (and other API clients) from creating devices
in sites when sites functionality is disabled for the organization.
Fixes issue where Terraform could bypass site restrictions via API.
Show the specific cloud poller name (e.g., dfw, lon) alongside "Cloud"
when a device is assigned to a cloud poller. Displays as "Cloud (dfw)"
instead of just "Cloud" for better visibility of which poller is being used.
When no agent is assigned, continues to show just "Cloud".
Display latency values with single decimal precision (e.g., 114.1ms)
for improved readability in graphs. This ensures consistency between
the raw chart data and tooltip formatting.
The previous commit (90e8964) introduced an incorrect LoggerBackends.add/2
call that prevented Oban from starting, causing all discovery jobs to remain
stuck in "available" state without being processed.
Root cause: LoggerBackends.add expects a tuple {module, id} as the first
argument, not the module alone with configuration options.
Fix: Use the correct API pattern:
1. Add backend with tuple identifier: {LoggerFileBackend, :file_log}
2. Configure separately with Logger.configure_backend/2
This resolves device discovery being completely broken - all background
jobs (discovery, polling, monitoring) now process correctly.
Critical fixes:
- Add [:safe] option to binary_to_term to prevent RCE attacks
- Implement whitelist validation for String.to_atom conversions
- Add input validation before String.to_existing_atom usage
Changes:
- MIB compiler and cache: Use safe binary deserialization
- SNMP contexts: Whitelist protocol, device type, and source atoms
- API controllers: Validate error message keys before atom conversion
- Reduce function nesting to comply with Credo standards
All 6,145 tests passing with zero Credo issues.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Replace Docker-based builds with Nix flakes for reproducible builds.
Changes:
- Use nixos/nix:latest image with `nix` tag requirement
- Build Docker image with `nix build .#dockerImage`
- Add optional Cachix integration for binary caching
- Load image with `docker load < result`
- Keep existing deploy stage unchanged
Benefits:
- Reproducible builds across all environments
- Binary caching via Cachix (faster subsequent builds)
- All dependencies pinned in flake.lock
- Image size similar to Docker builds (~507 MB compressed)
Requirements:
- GitLab Runner with Nix installed and `nix` tag
- Docker socket mounted for loading/pushing images
- Optional: CACHIX_AUTH_TOKEN for binary caching
See docs/gitlab-runner-nix-setup.md for runner setup instructions.
The old Docker-based config is available in git history if rollback needed.
Nix build fixes:
- Override beamPackages.elixir to use 1.19.5 instead of default 1.18.4
- Add explicit buildPhase to compile dependencies before application
- Remove postInstall phase (Mix automatically includes priv/ in release)
- Set HEX_OFFLINE=false to allow dependency fetching in Nix sandbox
Docker image build now succeeds (~507 MB compressed).
Also remove .claude/settings.local.json from git tracking (keep local).
File remains in working directory but is now properly ignored.
Add git and cacert to nativeBuildInputs and set SSL certificate
environment variables. This fixes the SSL certificate verification
error when fetching heroicons from GitHub.
Override configurePhase to explicitly run 'mix deps.get' before
compilation. This ensures hex dependencies (like ecto) are available
when vendored path dependencies (oban_met) are compiled.
The postUnpack step was deleting deps directories which may have
been removing vendored dependencies. cleanSourceWith already handles
source filtering, so postUnpack is redundant and potentially harmful.
Remove buggy conditional that was always true. Now the filter
correctly excludes only build artifacts while including all
source files including vendor directory.
Replace sourceFilesBySuffices with cleanSourceWith to ensure
vendor directory and all dependencies are properly included.
This fixes the oban_met compilation error.
Find the erts include path dynamically instead of relying on
erlang.version which may not match the actual erts version.
This fixes 'erl_nif.h' file not found error.
Instead of hardcoding the version in the path, find the release
directory dynamically. This fixes the 'undefined variable version'
error during Docker image build.
Ignore local Nix development directories and files:
- .nix-mix/ - Mix cache
- .nix-hex/ - Hex cache
- .nix-postgres/ - PostgreSQL data directory
- .nix-redis/ - Redis data directory
- .nix-services-started - Service flag file
These are created by the Nix dev shell and should not be
committed to the repository.
- Check if PostgreSQL/Redis already running before starting
- Install Hex and Rebar3 non-interactively on shell entry
- Only run migrations if deps directory exists
- Better error messages with log file locations
This fixes issues when:
- Port 5432 is already in use (existing PostgreSQL)
- Port 6379 is already in use (existing Redis)
- Mix prompts for Hex installation interactively
Add Nix overlay to pin specific versions matching .tool-versions:
- Erlang: OTP 28.3 (via erlang_28)
- Elixir: 1.19.5-otp-28 (via beam.packages.erlang_28.elixir_1_19)
This ensures the Nix build environment matches the asdf-based
development setup exactly, providing consistency across all
development and build environments.
Use stdenv.isLinux instead of lib.isLinux to properly check if
inotify-tools should be included. This fixes the undefined variable
error when evaluating the flake on macOS.
- Make inotify-tools Linux-only (not available on Darwin)
- Replace nixfmt-rfc-style with nixfmt (deprecated warning)
- Add flake.lock for reproducible builds
Changes:
- nix/shell.nix: Conditionally include inotify-tools only on Linux
- flake.nix: Use pkgs.nixfmt instead of pkgs.nixfmt-rfc-style
- flake.lock: Lock dependencies for reproducibility
- Move last_discovery_at update to after successful processing so failed
discoveries are retried on the next poll cycle instead of being blocked
for 24 hours
- Add disabled assignment check to agent polling target query so devices
with enabled=false assignments don't leak through site/org cascade
- Wrap spawned polling data processing in try/rescue so crashes are logged
instead of silently disappearing
Comprehensive documentation covering architecture, usage, technical
details, and future enhancements for the job monitoring dashboard.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Added integration tests for real-time updates and metrics display.
Also fixed dialyzer warnings in worker event broadcasting by removing
unreachable error handling branches.
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
Add Job Monitoring card to admin dashboard with:
- Hero icon for visual identification
- Brief description of monitoring capabilities
- Link to /admin/monitoring dashboard
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
