Django QueryWatch
Automated database-performance analysis for Django.
Source code and issue tracking: Atiqumer/django-querywatch.
QueryWatch observes database executions made by a Django request or test, groups query patterns, and turns suspicious behavior into explainable findings. It is a local-development and test/CI tool; it does not modify queries, create database tables, or send data to an external service.
Status: pre-alpha. The public API may change before 1.0.
Why QueryWatch?
Query viewers show what ran. QueryWatch focuses on what may deserve attention:
- queries at or above a configurable duration threshold;
- repeated query patterns;
- possible N+1 patterns with distinct parameters; and
- query budgets for tests.
Findings are rule-based and intentionally conservative. In particular, an N+1 finding is described as possible because SQL alone cannot prove which Python source line or Django relationship caused it.
Compatibility
QueryWatch targets Python 3.12–3.14 and Django 5.2–6.1. SQLite is used for the current test suite. PostgreSQL and MySQL compatibility verification is planned before a stable release.
Installation
After the package is published to PyPI:
python -m pip install django-query-sentinel
For development from a checkout, see Development.
Quick start
# settings.py
INSTALLED_APPS = [
# ...
"querywatch",
]
MIDDLEWARE = [
# ...
"querywatch.middleware.QueryWatchMiddleware",
]
The middleware logs a console report through the querywatch logger after
each synchronous request. Set QUERYWATCH_OUTPUT = "none" when you only want
to use the testing API.
Configuration
# settings.py
QUERYWATCH_ENABLED = True
QUERYWATCH_OUTPUT = "console" # "console", "json", or "none"
QUERYWATCH_SLOW_QUERY_MS = 100
QUERYWATCH_DUPLICATE_THRESHOLD = 3
QUERYWATCH_NPLUS1_THRESHOLD = 3
QUERYWATCH_MAX_QUERIES = None # Request-level finding; disabled
QUERYWATCH_MAX_QUERIES causes an error-level finding in request reports. It
does not change an HTTP response. To fail a test, use assert_queries().
Reports
Console output is intentionally dependency-free:
QueryWatch
────────────────────────────────
Queries: 27
Total DB time: 84.3ms
Slow queries: 2
Duplicate groups: 4
Findings
────────────────────────────────
WARNING Possible N+1 query pattern
A query pattern was executed repeatedly with different parameters.
Set QUERYWATCH_OUTPUT = "json" to log a machine-readable JSON object with a
summary and findings. Reports include structural evidence, timing values, and
fingerprints; they do not include query parameters.
Query budgets
from querywatch.testing import assert_queries
def test_dashboard(client):
with assert_queries(max_queries=10):
response = client.get("/dashboard/")
assert response.status_code == 200
If the budget is exceeded, QueryWatchError exposes actual_queries,
max_queries, exceeded_by, findings, and report for programmatic
assertions.
Detectors
| Detector | Default | Severity |
|---|---|---|
| Slow query | QUERYWATCH_SLOW_QUERY_MS = 100 |
warning |
| Duplicate query | QUERYWATCH_DUPLICATE_THRESHOLD = 3 |
info |
| Possible N+1 | QUERYWATCH_NPLUS1_THRESHOLD = 3 |
warning |
| Query count | disabled (QUERYWATCH_MAX_QUERIES = None) |
error |
Privacy and safety
- Captured query data exists only in memory for the active scope.
- QueryWatch does not create models, migrations, caches, queues, or services.
- QueryWatch does not re-execute captured SQL and never runs
EXPLAIN ANALYZE. - Query parameters are not included in console or JSON reports.
- QueryWatch does not change a QuerySet, transaction, SQL statement, or result.
Limitations
- Middleware instrumentation is synchronous. Native async instrumentation is not claimed in v0.1.
- SQL normalization is deliberately conservative; comments and PostgreSQL dollar-quoted strings are left unchanged.
- Possible-N+1 detection is heuristic and does not infer exact relationship names or source locations.
- This is not a production-monitoring service or a historical metrics store.
Development
Requires Python 3.12+.
python -m pip install -e ".[dev]"
python -m ruff check .
python -m pytest
Run the complete test suite before submitting a change. Tests use an isolated SQLite Django test database and require no network access.
Packaging verification
Build both distribution formats locally:
python -m build
This creates a wheel and source distribution in dist/. Verify the wheel in
a fresh virtual environment before publishing:
python -m venv .wheel-test
.\.wheel-test\Scripts\python -m pip install .\dist\django_query_sentinel-0.1.1-py3-none-any.whl
.\.wheel-test\Scripts\python -c "import querywatch; print(querywatch.__name__)"
Replace the wheel filename if the package version changes. Do not publish until the wheel installs and imports successfully.
For the tag-based Trusted Publishing release process, see RELEASING.md.
Contributing
See CONTRIBUTING.md. Please keep changes small, add tests for behavior changes, and avoid new runtime dependencies unless there is a concrete need.
License
QueryWatch is released under the MIT License.
Metadata
Release files for django-query-sentinel 0.1.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | Actions |
|---|---|---|---|---|
| django_query_sentinel-0.1.1.tar.gz | 29.5 kB | With provenance | Details for django_query_sentinel-0.1.1.tar.gz |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| django_query_sentinel-0.1.1-py3-none-any.whl | Python 3 | none | any | Details for django_query_sentinel-0.1.1-py3-none-any.whl |
Total release size: 47.0 kB
Release files / django_query_sentinel-0.1.1.tar.gz
| Download URL | django_query_sentinel-0.1.1.tar.gz |
|---|---|
| Size | 29.5 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
33b7861f0901d4804c51a43ef02c85239e947b91915148cf3fe49f14c855c82a
|
|
BLAKE2b-256 checksum How to use checksums |
07b190d346d4e4f01d0bbbd4c0a2d886a437a6add49e39e2e61df106169f239a
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 26, 2026.
Transparency logRelease files / django_query_sentinel-0.1.1-py3-none-any.whl
| Download URL | django_query_sentinel-0.1.1-py3-none-any.whl |
|---|---|
| Size | 17.5 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
559ac021013993cf1c198d7f603cd1af5038b49cac5f1330ed205856e516abac
|
|
BLAKE2b-256 checksum How to use checksums |
fcae12c0a225a5dcf1779c8b1a746c239b09b124abe9256e4b8dc42378538d15
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 26, 2026.
Transparency logSecurity for django-query-sentinel 0.1.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
Both files in this release have been uploaded with provenance metadata.
Attestation for django_query_sentinel-0.1.1.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Sep 26, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.