Imported from victoriametrics/skills (plugins/query/skills/victorialogs-query/SKILL.md). Install upstream with npx skills add victoriametrics/skills --skill victorialogs-query. Copyright stays with the author.

VictoriaLogs Query

Query VictoriaLogs HTTP API directly via curl. Covers log search, stats queries, field/stream discovery, hits analysis, and facets.

Environment

# $VM_LOGS_URL - base URL
#   Example: export VM_LOGS_URL="https://vlselect.example.com"
# $VM_CURL_CONFIG - curl config file with auth header (set for remote, unset for local)
#   Remote: export VM_CURL_CONFIG="$HOME/.config/victoriametrics/curl.conf"
#   Local:  leave unset (defaults to /dev/null - no auth)

Auth Pattern

All curl commands load auth from a curl config file:

curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  "$VM_LOGS_URL/select/logsql/query?query=*&start=2026-03-07T00:00:00Z&limit=10"

When VM_CURL_CONFIG is unset, curl reads /dev/null and sends no auth header. When set, it must point to a mode-0600 curl config file containing header = "Authorization: Bearer <token>". Never print its contents.

Critical Rules

• ALWAYS pass start on ALL VictoriaLogs endpoints — omitting it scans ALL stored data (extremely expensive). Not technically required by the API, but omitting it is almost never what you want.
• stats_query uses time (ALWAYS pass it explicitly; do not rely on a default), stats_query_range uses start/end/step
• step is required for hits endpoint
• /select/logsql/query returns JSON Lines (one JSON object per line), NOT a JSON array
• LogsQL queries with special characters MUST be URL-encoded (use --data-urlencode for POST)
• stats_query and stats_query_range queries MUST contain a | stats pipe

Core Endpoints

Log Query

# Basic query (last hour, limit 100)
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query={namespace="myapp"} error' \
  "$VM_LOGS_URL/select/logsql/query?start=2026-03-07T00:00:00Z&limit=100"

# With time range and field selection
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query={namespace="myapp"} error' \
  "$VM_LOGS_URL/select/logsql/query?start=2026-03-07T00:00:00Z&end=2026-03-07T12:00:00Z&limit=50&fields=_time,_msg,level"

Parameters: query (required), start (required, RFC3339), end, limit, offset, fields (comma-separated), timeout

Response: JSON Lines (one JSON object per line). Pipe through jq -s . to collect into array, or process line-by-line.

Stats Query (Instant)

# Count errors by level at a point in time
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query={namespace="myapp"} | stats by (level) count() as total' \
  "$VM_LOGS_URL/select/logsql/stats_query?time=2026-03-07T09:00:00Z" | jq .

Parameters: query (required, must contain | stats pipe), time (required, RFC3339), start (optional, alternative to time).

Response: Prometheus-compatible JSON format.

Stats Query Range

# Error count over time with 1h steps
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query={namespace="myapp"} error | stats count() as total' \
  "$VM_LOGS_URL/select/logsql/stats_query_range?start=2026-03-07T00:00:00Z&end=2026-03-07T12:00:00Z&step=1h" | jq .

Parameters: query (required, must contain | stats pipe), start (required), end, step (e.g., 1h, 5m)

Response: Prometheus matrix format.

Hits (Log Volume)

# Log volume over time
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query={namespace="myapp"}' \
  "$VM_LOGS_URL/select/logsql/hits?start=2026-03-07T00:00:00Z&end=2026-03-07T12:00:00Z&step=1h" | jq .

Parameters: query (required), start (required), end, step (required), field (optional, group by field)

Facets (Best Discovery Tool)

# Discover field value distributions
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query={namespace="myapp"}' \
  "$VM_LOGS_URL/select/logsql/facets?start=2026-03-07T00:00:00Z&end=2026-03-07T12:00:00Z" | jq .

Parameters: query (required), start (required), end. Returns most frequent values for ALL fields in one call.

Discovery Endpoints

Field Names and Values

# Discover non-stream field names
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query={namespace="myapp"}' \
  "$VM_LOGS_URL/select/logsql/field_names?start=2026-03-07T00:00:00Z" | jq .

# Get values for a specific field
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query={namespace="myapp"}' \
  "$VM_LOGS_URL/select/logsql/field_values?start=2026-03-07T00:00:00Z&field=level&limit=20" | jq .

Stream Fields

# Discover stream field names (e.g., namespace, pod)
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query=*' \
  "$VM_LOGS_URL/select/logsql/stream_field_names?start=2026-03-07T00:00:00Z" | jq .

# Get values for a stream field
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query=*' \
  "$VM_LOGS_URL/select/logsql/stream_field_values?start=2026-03-07T00:00:00Z&field=namespace" | jq .

Streams and Stream IDs

# List log stream identifiers
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query={namespace="myapp"}' \
  "$VM_LOGS_URL/select/logsql/streams?start=2026-03-07T00:00:00Z&limit=20" | jq .

# List stream IDs
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query={namespace="myapp"}' \
  "$VM_LOGS_URL/select/logsql/stream_ids?start=2026-03-07T00:00:00Z" | jq .

LogsQL Quick Reference

LogsQL is space-separated (AND by default). Pipes use |.

# Stream filter (fast, use for namespace/pod selection)
{namespace="myapp"}

# Word filter (matches full words in _msg)
{namespace="myapp"} error

# Multiple words (AND)
{namespace="myapp"} error timeout

# OR filter
{namespace="myapp"} (error OR warning)

# Regex filter on _msg
{namespace="myapp"} ~"err|warn"

# Case-insensitive regex
{namespace="myapp"} ~"(?i)error"

# Field-specific filter
{namespace="myapp"} level:error

# Time filter (alternative to API start/end params)
{namespace="myapp"} _time:1h error

# Negation
{namespace="myapp"} error -"expected error"

# Stats pipe
{namespace="myapp"} _time:1h | stats by (level) count() as total

# Sort pipe
{namespace="myapp"} error | sort by (_time)

# Top pipe
{namespace="myapp"} _time:1h | top 10 (level)

Common mistakes:

• | grep does NOT exist. Use word filters or ~"regex".
• | filter is a valid pipe but ONLY after | stats for filtering aggregated results.
• Time ranges go in API start/end params OR _time: filter, NOT both.
• Stream field names depend on your ingestion config. ALWAYS discover them first.
• Searching "error" without filtering vmselect: vmselect logs contain PromQL text with "error" — add -"vm_slow_query_stats" to exclude.

Timestamp Format

All times use RFC3339 format: 2026-03-07T09:00:00Z. Unix timestamps are NOT supported by VictoriaLogs.

Response Parsing (jq)

# Collect JSON Lines into array
... | jq -s .

# Extract message and timestamp from query results
... | jq -s '.[] | {time: ._time, msg: ._msg}'

# Count results from JSON Lines
... | jq -s 'length'

# Extract specific fields
... | jq -s '.[] | {time: ._time, level: .level, msg: ._msg}'

# Stats query — extract metric values
... | jq '.data.result[] | {metric: .metric, value: .value[1]}'

# Stats range query — extract time series
... | jq '.data.result[] | {metric: .metric, values: .values}'

# Facets — list fields and top values
... | jq '.[] | {field: .name, values: [.values[] | .value]}'

Common Patterns

# Quick error check for a namespace (last hour)
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query={namespace="myapp"} error' \
  "$VM_LOGS_URL/select/logsql/query?start=$(date -u -d '1 hour ago' +%Y-%m-%dT%H:%M:%SZ)&limit=20"

# Error rate over time
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query={namespace="myapp"} error | stats count() as errors' \
  "$VM_LOGS_URL/select/logsql/stats_query_range?start=2026-03-07T00:00:00Z&step=1h" | jq .

# Discover all namespaces with logs
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query=*' \
  "$VM_LOGS_URL/select/logsql/stream_field_values?start=2026-03-07T00:00:00Z&field=namespace" | jq .

# Search by trace ID in logs
curl -q --config "${VM_CURL_CONFIG:-/dev/null}" -s \
  --data-urlencode 'query=trace_id:"abc123def456"' \
  "$VM_LOGS_URL/select/logsql/query?start=2026-03-07T00:00:00Z&limit=50"

Environment Switching

# Check current environment
echo "VM_LOGS_URL: $VM_LOGS_URL"
echo "VM_CURL_CONFIG: ${VM_CURL_CONFIG:-(unset - no auth)}"

Important Notes

• POST is preferred for queries with special characters — use --data-urlencode to avoid URL encoding issues
• ALL endpoints accept both GET and POST with application/x-www-form-urlencoded
• field parameter is required for field_values and stream_field_values endpoints
• facets is the best single-call discovery tool — returns all field distributions at once
• Do NOT confuse stats_query (instant, uses time) with stats_query_range (range, uses start/end/step)
• For full endpoint details, parameters, and response formats, see references/api-reference.md