Runs · gharchive · ← run 4074 · run 4083 →
Run 4076 succeeded
Log in to rerunStatusWhere the run is: queued, running, then succeeded, failed or cancelled. Refreshed every 5 s while it goes on. Docs ↗
- DurationHow long the run took, from its start to its end, or to now while it goes on. Hover for the times.
- 1 min 45 s
- CreatedWhen the run was queued: by the schedule, a sensor, a backfill, a fill or a person.
- 2026-11-08T12:40:13Z
- StartedWhen its process started, once its turn in the queue came and its placement had room. Empty while it waits.
- 2026-11-08T12:40:13Z
- EndedWhen it finished, whatever the outcome. Empty while it goes on.
- 2026-11-08T12:41:58Z
- TriesTimes the run's process was started: 1 is normal; more means retries, a lost instance or a resume.
- 1
Partitions & triggerWhat this run does, and what made it start. Docs ↗
CodeThe code this run ran, at the commit its binary was built from. Hover a name to see its source; click it to open the file on its host. Docs ↗
- FunctionsThe pipeline, its steps in the order they run, and the dista tasks they fan out to, each linked to its line at the commit this run ran. Docs ↗
- gharchive · load_day · index_hour · hour_task · merge_repo_hour
- SourceThe commit its binary was built from (-D ORCA_COMMIT), in which repository, and the image it ran in, if any. Docs ↗
- ac0c766d3f15 of https://github.com/magmalake/mojo-census
- Driver logWhat the run's process printed: orca's lines and its steps' output, as the scheduler kept it. Each attempt under Steps has its own log too.
- driver.log
Code hash 0xe6cdc4870d3f3535A hash of what the binary says about itself: its steps, partitions and schedule. It changes when they do, not on every build. Docs ↗
ResourcesWhat this run's own processes used: its process and the ones it started, dista's local workers too, measured by the scheduler or worker that started it. Not the rest of the machine: the machine's numbers, where they are recorded, are in their own card further down. With a request, what it asked for and was granted too. Docs ↗
- Peak memoryThe most memory its processes held at once: the run's process and what it started (dista's local workers too), sampled every second by the scheduler or worker that started it, and at the end the largest single process. Not measured on EC2 or Kubernetes. Docs ↗
- 665 MB
- CPU usedThe CPU time its processes used, user and system, and the most cores they kept busy over one sample's second. macOS gives only the time, and only for a process that made one attempt. Docs ↗
- 146.6 s, up to 3.0 cores at once
SettingsWhat the run was set up with, orca.toml and the flags applied: as its start recorded them, or else as the pipeline is registered. The pipeline's Config card says where each value came from. Docs ↗
As the pipeline is registered. Where each value came from
- PlacedWhere it ran: local (beside the scheduler, on its machine), control (on the control host, within its limits), ec2 (instances of its own), or a remote worker's pool. Docs ↗
- local
- Max runsThe most runs of this pipeline the scheduler starts at once (max_runs); the others wait in the queue.
- 1
- SpotWhether its EC2 instances are Spot: cheaper, and AWS can take them back.
- false
- EnvironmentThe variables orca.toml's env and the flags set in the run's environment.
DISTA_WORKER_SLOTS=2GHARCHIVE_EVENTS=offGHARCHIVE_PARALLEL=2
StepsEach step of each partition. Table: its status, attempts, times, what its latest attempt used, and its log. Timeline: the attempts on the clock, a row per step. Graph: the steps in the order they run, in this run's colours. Docs ↗succeeded
| StepA step of the pipeline: a function that runs once per partition (mapped) or once over all of them (gather). | PartitionThe partition this row is for; empty for a gather step, which runs once over all of them. | StatusWhere it stands. Hover a status for what it means. | AttemptsHow many times the step ran in this run: more than 1 means an attempt failed and it was retried. | StartedWhen it started; empty while it waits. | EndedWhen it finished, whatever the outcome. | PeakThe most memory the step's latest attempt held at once; its page has each attempt's. | CPUCPU used: for a step, its latest attempt's user and system seconds; for a machine, its busiest moment as a share of all cores. | LogIts output: an attempt's, an instance's or the run's process's. | ErrorWhy it failed, as the step or orca said. |
|---|---|---|---|---|---|---|---|---|---|
| load_day pipelines/gharchive.mojo:135 | 2025-06-27 | succeeded | 1 | 2026-11-08T12:40:13Z | 2026-11-08T12:41:57Z | 665 MB | 146.6 s | attempt 1 |
MachinesSamples of the whole machine, every few seconds while the run ran (ORCA_SAMPLE_SECONDS, 5 by default): its CPU, memory, fullest disk, disk io and network, by a sampler the runner starts beside the run's process. They count everything on that machine, not just this run. What the run's own processes used is in Resources, above. The steps lane lines a busy stretch up with the step it came from. A run on a remote worker has none: the worker measures only its processes. Docs ↗
The whole machine, while this run ran: the scheduler's machine (the local runner). Not this run's own use; that is in Resources.
A shared machine: 5 other runs ran on it at the same time, so these numbers are theirs too.
user + system iowait steal memory fullest disk disk io network warning
| InstanceWhere the samples were taken: local (the scheduler's machine), control, an EC2 head or worker, or a pod. | SamplesHow many samples the sampler took. | CPUCPU used: for a step, its latest attempt's user and system seconds; for a machine, its busiest moment as a share of all cores. | Busiest coreThe busiest single core's use at its highest. | StealCPU time the hypervisor gave to other machines. | MemoryMemory used at its highest, of the machine's total. | DiskThe fullest disk the run writes to, at its fullest. | Disk ioDisk reads and writes, at their highest rate. | NetworkNetwork traffic in and out, at its highest rate. | WarningsStretches with memory or a disk over 90%, or CPU steal over 20%: what slows or fails most runs. |
|---|---|---|---|---|---|---|---|---|---|
local | 22 | 100% | 100% | 0% | 2.0 GB of 8.1 GB (24%) | 9.7 GB of 80 GB (12%) | 204 kB/s | 112 MB/s |
Timeline
The timeline loads as it comes into view.
Wheel to zoom, drag to zoom to a span of time, hover for a span's details.
SnapshotsThe table snapshots each step read, and the ones it wrote (committed). A dataset's snapshot links to its lineage: upstream, what it was written from, and downstream, what was written from it. Docs ↗
| StepA step of the pipeline: a function that runs once per partition (mapped) or once over all of them (gather). | PartitionThe partition this row is for; empty for a gather step, which runs once over all of them. | Read or wroteread: the step read this snapshot. wrote: it committed it, a new snapshot of the table. | TableThe table's name in the warehouse. | SnapshotA snapshot of a table: the table as one commit left it. Click for its lineage. |
|---|---|---|---|---|
| load_day | 2025-06-27 | wrote | gharchive.candidates | 7320659257681652334 |
| load_day | 2025-06-27 | wrote | gharchive.repo_hour | 3760277588636738099 |
| load_day | 2025-06-27 | wrote | gharchive.releases | 1393842515952960652 |