Observe
Get insights on every dbt™ schedule
Turn Bolt run history into insights on schedules, models, sources, and tests. See which schedules fail most and how long they take, and which models are slow or fragile. See which sources went stale and which tests keep failing.
Fivetran1m 08s
Airbyte52s
dlt41s
Run DinoAI Background Agent
Run dbt™ Command
Run Python Script
Run pre-commit
Run Elementary
Run MonteCarlo CLI
Refresh PowerBI Dataset
Refresh Tableau WorkbookRefresh Tableau Data Source
Run Lightdash
Run Hex Project
Sync Fivetran Connector
Sync Airbyte Connection
Sync Hightouch
Run BigQuery Scheduled Query
Control GCP Datastream
Trigger Airflow DAG
Run Matillion Pipeline
Run Azure Data Factory Pipeline
Run AWS Step Functions
Trigger GitHub Actions
Trigger Zapier
Run AWS Glue
Invoke AWS Lambda
Run AWS ECS Task
Run AWS SageMaker Pipeline
Run GCP Dataflow
Invoke GCP Cloud Function
Run GCP Cloud Run Job
Run GCP Dataproc Job
Push to DataHubToggle Paradime Bolt ScheduleRefresh Paradime Catalog
dbt build4m 12s
Python script37s
DinoAI agent1m 20s
Tableau44s
PowerBI31s
Hightouch18s
Slack#data-alerts
MS TeamsData platform
PagerDutyIncidentsWhen a run fails, a DinoAI agent diagnoses it in the failure thread, implements a fix, and opens a pull request. It never merges on its own.
DinoAI
GitHubschedules:- name: nightly runenvironment: productiongit_branch: mainschedule: "0 4 * * *"timezone: UTCcommands:- paradime run fivetran-sync- paradime run airbyte-sync- dbt build- python scripts/score.py- paradime run tableau-refresh- paradime run hightouch-syncnotifications:slack_channels:- channel: '#data-alerts'events: [failed, sla]self_healing:enabled: trueagent_name: bolt-pipeline-healer
04:02:11 Running with dbt=1.9.404:02:13 Found 148 models, 312 data tests, 12 sources04:02:14 Concurrency: 4 threads (target='prod')04:02:16 1 of 460 START sql table model analytics.stg_orders04:02:19 1 of 460 OK created sql table model analytics.stg_orders [SUCCESS 1 in 2.41s]04:06:21 460 of 460 PASS not_null_fct_orders_order_id [PASS in 0.38s]04:06:23 Finished running 148 models, 312 data tests in 4m 12s04:06:23 Done. PASS=460 WARN=0 ERROR=0 SKIP=0 TOTAL=460
5 downstream nodes impacted- order_total+ order_amount
fct_orders.order_totaldbt™ modelmart_revenue.total_revenuedbt™ modelRevenue overviewTableau dashboard
Exec KPIsLooker dashboard
Sales pipelineThoughtSpot liveboardFeatures
Everything dbt™ monitoring gives you
Insights from your production runs so you can drive your roadmap what needs fixing
Executed Nodes
See what was executed in each dbt™ run and deep dive into models, and tests that failed.
Model timeline
Find which are your long-running and bottleneck models and how you can make your pipeline run faster.
Console logs
Get access to the logs of every command that was run and switch to the DinoAI summary to see what's broken and then fix with DinoAI agents.
Command artifacts
View every SQL and DDL statement that was executed on the warehouse or download logs for further analysis.
Schedule durations
Use the schedules dashboard to deep dive into your longest schedules, root cause and DinoAI recommended fixes to improve speed and reduce cost.
Execution time history
Find outliers in your schedule execution so that you can find what code changes drove that.
Workflows
What you answer with pipeline observability
The questions a data team gets asked, and where the answer lives.
Reliability
Which schedules keep failing, and why?
The Schedules dashboard ranks error counts and runtime by schedule. Click through to the run, read the DinoAI summary of the failed command, and hand it to Fix with DinoAI.
Performance
Which model is the bottleneck?
The Models dashboard shows the longest-running models and their runtime trend. The run's model timeline shows where the critical path waits. Convert the slow model to incremental from the Code IDE.
Freshness
Did yesterday's data actually land?
The Sources dashboard shows every source's freshness state and last loaded time. Add a source freshness command to the schedule, and the run itself checks freshness before it continues.
Quality
Which tests fail most, on which models?
The Tests dashboard lists non-passing tests with their models and impacted row counts, and the models with the most warnings. Prioritize the refactor from there.
SLA
Will the 6am run going to miss SLA?
Set an SLA on the schedule and route the breach to Slack, Teams, email, or your incident tool. The templated message carries the run link, branch, commit, and duration.
Root cause
What ran when the pipeline failed?
Open the run, switch between pipeline, executed nodes, and model timeline, search the logs, and download the compiled SQL. Retry from failure once you know.
Join industry leaders winning with DinoAI agents
Data engineers and analysts at the most AI-native startups and enterprises use Paradime's DinoAI agents everyday to become market leaders.