apache/druid · error · SamplerException

Failed to sample data: %s

Error message

Failed to sample data: %s

What it means

InputSourceSampler wraps any exception raised while reading and parsing sampled input data into a SamplerException with this message. Sampling runs a short-lived reader over the input source (e.g. for the data loader or ingestion spec preview); failures there — bad input source config, unreadable storage, malformed records — surface through this catch-all. The original exception is attached as the cause and its message is embedded in the entity.

Source

Thrown at indexing-service/src/main/java/org/apache/druid/indexing/overlord/sampler/InputSourceSampler.java:290

            );
          }
        }

        return new SamplerResponse(
            numRowsRead,
            numRowsIndexed,
            logicalDimensionSchemas,
            physicalDimensionSchemas,
            signatureBuilder.build(),
            responseRows.stream()
                        .filter(Objects::nonNull)
                        .filter(x -> x.getParsed() != null || x.isUnparseable() != null)
                        .collect(Collectors.toList())
        );
      }
    }
    catch (Exception e) {
      throw new SamplerException(e, "Failed to sample data: %s", e.getMessage());
    }
  }

  private InputSourceReader buildReader(
      SamplerConfig samplerConfig,
      DataSchema dataSchema,
      InputSource inputSource,
      @Nullable InputFormat inputFormat,
      File tempDir
  )
  {
    final InputRowSchema inputRowSchema = InputRowSchemas.fromDataSchema(dataSchema);

    InputSourceReader reader = inputSource.reader(inputRowSchema, inputFormat, tempDir);

    if (samplerConfig.getTimeoutMs() > 0) {
      reader = new TimedShutoffInputSourceReader(reader, DateTimes.nowUtc().plusMillis(samplerConfig.getTimeoutMs()));
    }

View on GitHub (pinned to 9b90983fd2)

Solutions

1. Read e.getMessage() / the exception cause in the response — it names the real failure
2. Fix the inputSource config (paths, URIs, connection settings, credentials) and retry the sample
3. Align the parser/inputFormat with the actual data format before sampling
4. Reduce samplerConfig maxRows/targetPartitionSize if huge rows trigger the flush threshold error

Example fix

// before
{"type":"s3","uris":["s3://bucket/wik.json"]}  // no credentials configured
// after
{"type":"s3","properties":{"accessKeyId":"...","secretAccessKey":"..."},"uris":["s3://bucket/wik.json"]}

// sanity-check the input source config before sampling
if (!spec.inputSource || !spec.parser && !spec.inputFormat) throw new Error('inputSource and inputFormat required');

function isSampleable(spec) { return spec != null && spec.inputSource != null && spec.inputSource.type != null && (spec.inputFormat != null || spec.parser != null); }

try { await sample(spec); } catch (e) { if (/Failed to sample data/.test(e.message)) { inspectCause(e); fixInputSourceConfig(); } else throw e; }

When it happens

Trigger: POST /druid/indexer/v1/sampler (or via the web-console data loader preview) where opening/reading the InputSource throws: wrong S3/GCS/local path or credentials, unsupported input format, parser/schema mismatch, or the sampler's flush threshold is exceeded.

Common situations: Previewing a Kafka/Kinesis topic with misconfigured bootstrap servers or no read permission; sampling a file whose format does not match the parseSpec; sampling data whose rows fail every parse attempt plus an internal reader bug; timeout reading cold storage.

Understand the failure class

Background: "API request failed": what wrapped HTTP errors from external APIs mean and how to find the real cause — this error's family across 29 libraries.

Related errors

• No task information found for task with id: [%s]
• Cannot find any supervisor with id: [%s]
• authResult.getErrorMessage()
• Cannot find any task with id: [%s]
• authResult.getErrorMessage()

AI-assisted analysis of apache/druid@9b90983fd2 (2026-09-07). Data as JSON: /api/errors/e18e5af589ac3d1b. Report an issue: GitHub.