ClearActiveAlerts triggers an async save to disk, which can race with
LoadActiveAlerts reading the file. The test now clears the in-memory
map directly without triggering the async save.
- Add AI service with Anthropic, OpenAI, and Ollama providers
- Add AI chat UI component with streaming responses
- Add AI settings page for configuration
- Add agent exec framework for command execution
- Add API endpoints for AI chat and configuration
Host agent was including Docker overlay2 mounts from TrueNAS SCALE's
.ix-apps directory in disk totals. These mounts inherit the ZFS pool's
AVAIL space, causing massively inflated storage numbers (e.g., 173 TB
per container overlay instead of actual usage).
Changes:
- Add /mnt/.ix-apps/docker/ to container overlay path exclusions
- Use ShouldSkipFilesystem() in host agent disk collection (was only
using ShouldIgnoreReadOnlyFilesystem() which missed container paths)
- Add test cases for TrueNAS overlay paths
Related to #718
- Rename checkFlapping to checkFlappingLocked to clarify lock contract
- Replace goto statements with structured control flow
- Wire up unused recordAlertFired/recordAlertResolved metric hooks
- Add trackingMapCleanup goroutine to prevent memory leaks from stale entries
- Tighten alert ID validation to alphanumeric + safe punctuation
- Fix history save error handling to properly manage backup lifecycle
- Add auto-migration for deprecated GroupingWindow field
- Refactor 300+ line UpdateConfig into focused helper functions
- Unify duplicate evaluateVMCondition/evaluateContainerCondition
- Add constants for magic numbers (thresholds, timing, flapping)
- Update tests to match new backup behavior
- Add unit tests for internal/buffer package
- Fix misleading "ring buffer" comment (it's a bounded FIFO queue)
- Remove unused BufferCapacity config field from both agents
- Rewrite flaky integration test to use polling instead of fixed sleeps
When new nodes are added to a Proxmox cluster after Pulse was
initially configured, they weren't showing up in Settings. The
existing "Refresh" button only triggered network discovery, not
cluster membership re-detection.
Changes:
- Add POST /api/config/nodes/{id}/refresh-cluster endpoint
- Add "Refresh" button in cluster node panel in Settings
- Re-detect cluster membership and update stored endpoints
Related to #799
The config package has a more robust IsPasswordHashed function that
handles truncated hashes. The auth package had a simpler duplicate
that was only used in tests. Removed the duplicate and its test
(already covered by config/config_utils_test.go).
Reduces deadcode findings from 7 to 6.
The isLoopback function in internal/hostagent/agent.go was unused in
production code - it was a duplicate of the same function in
internal/hostmetrics/collector.go which is actively used.
Removed the dead code along with its associated tests to reduce
maintenance burden and improve code clarity.
Refactor the test to avoid unreachable code after panic by
checking a flag set before the panic instead of after. This
resolves the go vet warning while maintaining test coverage.
Replace raw fmt.Printf calls with structured zerolog logging for
consistency with the rest of the codebase. This improves log
formatting and enables proper log level filtering.
Add tests for the remaining uncovered State methods:
- SyncGuestBackupTimes: Tests backup time sync from storage and PBS backups
- UpdateStorageBackupsForInstance: Tests storage backup updates with instance isolation
Improves internal/models coverage from 89.6% to 95.6%.
Add missing godoc comments to:
- BuildGuestKey in alerts/alerts.go
- GenerateMockData in mock/generator.go
- NewDockerUpdater, NewAURUpdater in updates/adapter_installsh.go
- NewMockUpdater in updates/mock_updater.go
Add missing godoc comments to:
- NewRateLimiter and Allow in ratelimit.go
- SnapshotSyncStatus in temperature_proxy.go
- NewClient and GetVersion in pkg/pmg/client.go
Cache err.Error() result in two locations:
- monitor.go: storage query retry logic (2x calls to 1)
- monitor_polling.go: storage timeout handling (2x calls to 1)
Removed two DEBUG log statements that were logging full nodes config
JSON at Info level. This was verbose and potentially exposed sensitive
configuration data (credentials, tokens) in logs.
strconv.Itoa is faster than fmt.Sprintf("%d", ...) because it doesn't
need to parse a format string. Changed 4 occurrences in monitoring
package where integers are converted to strings.
Move regex compilation from function bodies to package-level variables
to avoid recompilation when parsing version strings.
Affected regexes:
- semverRe: Matches semantic version format (X.Y.Z-prerelease+build)
- rcNumRe: Extracts RC number from prerelease strings
These are called multiple times during version comparison and update checks.
Move regex compilation from function bodies to package-level variables
to avoid recompilation on each RAID array check during polling cycles.
Affected regexes:
- mdDeviceRe: Matches md device names in /proc/mdstat
- slotRe: Matches device slot lines in mdadm --detail output
- speedRe: Extracts rebuild speed from /proc/mdstat
Storage that is disabled in Proxmox (Datacenter > Storage > enabled=no)
was incorrectly showing as "available" in Pulse. The issue was that
Enabled/Active fields defaulted to true and were never set to false
from the per-node API response.
Now the model correctly initializes Enabled/Active from the Proxmox
per-node storage API response, and the status determination prioritizes
checking the disabled state first.
Related to #796
- parseDetail: Remove len(parts) != 2 check after SplitN(line, ":", 2)
when we already verified line contains ":" (SplitN always returns 2)
Function improved from 96.4% to 98.1% coverage.
- checkOrigin: Remove redundant same-origin check at line 155 that was
already handled at line 116 (origin == requestOrigin)
Function improved from 95.1% to 97.4% coverage.
- Host.ToFrontend: Remove redundant h.DisplayName check after struct
initialization already set host.DisplayName = h.DisplayName
Function now at 100% coverage.
- errors.isRetryable: Convert final case to default (all ErrorType values covered)
- scheduler.SelectInterval: Remove bounds checks after mathematical computation
that guarantees target ∈ [min, max]
Both functions now at 100% coverage.
- firstForwardedValue: strings.Split always returns at least one element
- shouldRunBackupPoll: remaining is always >= 1 by math
- convertContainerDiskInfo: lowerLabel is never empty for non-rootfs
All three functions now at 100% coverage.
Add tests for sliding expiration session validation and no-auth
configured scenarios. These test explicit paths for better coverage
documentation even though they were already exercised indirectly.
Add comprehensive direct tests for the CheckProxyAuth function covering:
- Not configured (returns false)
- Invalid secret (returns false)
- Missing secret header (returns false)
- Valid secret without user header configured (returns true, admin)
- Missing user header when configured (returns false)
- Valid auth with username (returns true with username)
- Role checking with empty roles header (defaults to admin)
- Role checking with admin role present (returns admin=true)
- Role checking without admin role (returns admin=false)
- Custom role separator (comma instead of pipe)
- Role with whitespace (trimmed correctly)
Coverage: CheckProxyAuth 89.3% → 100%
Add comprehensive tests for the ValidateSession wrapper function covering:
- Non-existent token (returns false)
- Empty token (returns false)
- Valid token (returns true)
- Expired token (returns false)
The ValidateSession function is a simple wrapper around the SessionStore's
ValidateSession method, but having direct tests ensures the wrapper is
exercised and documents its expected behavior.
Coverage: ValidateSession 0% → 100%
Add comprehensive tests for the RequireAdmin middleware covering:
- No auth configured (allows access by design)
- API-only mode (rejects requests without token)
- Basic auth with invalid credentials
- Proxy auth with admin role (allowed)
- Proxy auth with non-admin role (forbidden)
- Proxy auth with invalid secret (unauthorized)
- Proxy auth without role header (defaults to admin)
- Proxy auth with custom role separator
- Proxy auth with spaces in roles (trimmed)
- Basic auth authenticated users (allowed as admin)
- JSON vs plain text error responses based on path/Accept header
Also improves CheckProxyAuth coverage as a side effect.
Coverage: RequireAdmin 20.8% → 87.5%
Coverage: CheckProxyAuth 0.0% → 89.3%
Coverage: API package 30.9% → 31.9%
Add comprehensive tests for the saveHistoryWithRetry function covering:
- Backup file creation from existing history
- Empty history serialization
- Single retry success
- Read-only directory failure with retries
- Concurrent saves with serialization via saveMu
- Snapshot isolation during save
Coverage: saveHistoryWithRetry 58.6% → 86.2%
Coverage: alerts package 87.4% → 87.8%
Tests for NotificationManager accessor and helper functions.
Covers queue retrieval, webhook delivery tracking, history trimming
to max 100 entries, and copy-on-read semantics. Notifications 56.6%→57.3%.
Add TestDispatchAlert with 8 test cases covering:
- Returns false when onAlert callback is nil
- Returns false when alert is nil
- Returns false when activation state is pending
- Returns false when activation state is snoozed
- Returns false for monitor-only alerts
- Dispatches synchronously when async is false
- Dispatches asynchronously when async is true
- Clones alert before dispatch
Alerts package coverage: 83.4%→83.5%
The quick-setup command for temperature monitoring was generating
--standalone --http-mode which is meant for Docker deployments. This
confused users trying to set up multi-server Proxmox monitoring.
Now uses --ctid which works for both local and remote Proxmox hosts.
The installer detects when the container doesn't exist locally and
installs in "host monitoring only" mode automatically.
If we can determine the actual CTID from the host proxy summary,
we use it; otherwise we show <PULSE_CTID> for the user to replace.
Related to #785
Fixed deadlock where CancelByAlertIDs held nq.mu.Lock() and then called
UpdateStatus() which also tried to acquire the same lock. Now uses
direct SQL while holding the lock.
Tests added for CancelByAlertIDs:
- No matching notifications (notification stays pending)
- Matching notification cancelled
- Multiple alerts with partial match (any match cancels)
Coverage: CancelByAlertIDs 65.7% -> 81.1%
Tests using NewManager() were sharing /etc/pulse/alerts, causing race
conditions when running in parallel. Added newTestManager(t) helper that
creates isolated temp directories for each test.
The test was failing after commit d6cbfc23 added security hardening
that requires authentication and trusted proxy configuration for
X-Forwarded-* headers to be read during public URL detection.
- Add API token authentication to the test request
- Configure 127.0.0.1 as trusted proxy for the test
- Add export_test.go with ResetTrustedProxyConfigForTests() to allow
external tests to reset the trusted proxy configuration
Add test for SaveNodesConfigAllowEmpty which permits explicit
deletion of all nodes. Document deadlock bug in saveNodesConfig
where empty config protection tries to call LoadNodesConfig
while holding write lock.
Add 6 tests for instance descriptor generation:
- No clients returns nil
- PVE only with sorted order
- PBS only
- PMG only
- All types combined
- Nil scheduler/stalenessTracker safety
Coverage: 60.7% → 62.5%
Add 5 tests to cover all branches:
- Not requested (ALLOW_ADMIN_BYPASS != "1")
- Enabled with PULSE_DEV=true
- Enabled with NODE_ENV=development
- Case-insensitive NODE_ENV check
- Declined when outside dev mode
Coverage: 40% → 100%
Host disk bars were showing virtual filesystems like tmpfs, /dev, /run,
/sys, and Docker overlay mounts. These clutter the UI and don't represent
meaningful disk usage.
Changed from `shouldIgnoreReadOnlyFilesystem` (read-only only) to the
full `fsfilters.ShouldSkipFilesystem` which also excludes:
- Virtual FS types: tmpfs, devtmpfs, sysfs, proc, cgroup, etc.
- Special mountpoints: /dev, /proc, /sys, /run, /var/lib/docker, /snap
- Network filesystems: fuse, nfs, cifs, etc.
Related to #790
UpdateVMsForInstance and UpdateContainersForInstance were replacing
guest data without preserving the LastBackup field that was populated
by SyncGuestBackupTimes. This caused backup indicators to always show
"no backup found" since the LastBackup would be wiped every time
guests were polled (which happens more frequently than backup polling).
Now both functions preserve LastBackup from existing data when the
incoming guest data has a zero value.
Related to #762
Tests empty string, invalid IP, and IP not matching CIDR for
isTrustedProxyIP. Also adds tests for GetClientIP empty RemoteAddr
and X-Real-IP fallback paths.
Tests WebSocket upgrade with POST method in demo mode, ensuring the
websocket check branch is exercised after the GET/HEAD/OPTIONS check.
Coverage: 90% to 100%.
- respondMissingScope: nil writer handling (80% to 100%)
- attachAPITokenRecord: nil record handling (80% to 100%)
- getAPITokenRecordFromRequest: wrong type and missing value (87.5% to 100%)
Add test case for URL containing control character (null byte)
which triggers url.Parse error and returns path unchanged.
Coverage improved from 88.9% to 100%.
