dash0-cli

name	dash0-cli
description	Query Dash0 logs from the CLI, manage profiles, and export CSVs with campaign/message fields. Use when users ask for Dash0 CLI commands, filters, profile selection, log export, or grouped campaign summaries.

Dash0 CLI

Use this skill when interacting with Dash0 from the terminal.

Quick Start

1. Check current profile, API URL, and dataset:

   dash0 config show
   

2. If needed, switch profile:

   dash0 config profiles list
   dash0 config profiles select prod
   

3. Query logs with experimental flag:

   dash0 logs query -X --from now-1h --to now
   

Query Syntax

• --filter is repeatable.
• Filter format is key operator value (not =).
• Common operator: is
• Example:

  dash0 logs query -X \
    --filter "service.name is lhgo" \
    --filter "otel.log.body is campaign.decide.outcome" \
    --from now-7d --to now
  

Export CSV

Use explicit columns so exported files are stable:

dash0 logs query -X \
  --from now-7d --to now \
  --filter "service.name is lhgo" \
  --filter "otel.log.body is campaign.decide.outcome" \
  --filter "reason is label_mismatch_disabled" \
  --column time \
  --column severity \
  --column service.name \
  --column body \
  --column reason \
  --column campaign_id \
  --column message_id \
  -o csv --limit 5000 > docs/export.csv

Important Notes

• Dataset names are case-sensitive. Example: default and Default can behave differently.
• If a query fails with 403 access denied, check profile/dataset with dash0 config show.
• --column and --filter can both be passed multiple times.
• Aliases are supported for columns: time, severity, body, etc.

Group-By Workaround

dash0 logs query does not provide SQL-style GROUP BY, so export first, then aggregate locally.

Group by campaign_id, include count and list of message ids, sorted by count descending:

awk -F',' '
NR==1 {
  for (i=1; i<=NF; i++) idx[$i]=i
  cid=idx["campaign_id"]; mid=idx["message_id"]
  if (!cid || !mid) {
    print "Missing campaign_id or message_id column" > "/dev/stderr"
    exit 1
  }
  next
}
{
  c=$cid; m=$mid
  if (c=="") next
  cnt[c]++
  if (msgs[c]=="") msgs[c]=m
  else msgs[c]=msgs[c]"|"m
}
END {
  for (c in cnt) print c "," cnt[c] "," msgs[c]
}
' docs/export.csv \
| sort -t, -k2,2nr \
| awk 'BEGIN { print "campaign_id,count,message_ids" } { print }' \
> docs/export_by_campaign.csv

Add URL column:

awk -F',' 'BEGIN {OFS=","} NR==1 {print $0,"campaign_url"; next} {print $0,"http://campaign-console.tailac0440.ts.net/campaigns/"$1}' \
  docs/export_by_campaign.csv > docs/export_by_campaign.tmp.csv \
  && mv docs/export_by_campaign.tmp.csv docs/export_by_campaign.csv

Response Style

When returning commands to users:

1. Give a copy/paste command first.
2. Include any assumptions (profile, dataset, time range).
3. If grouping is requested, provide both:
   • raw Dash0 query command
   • local aggregation command