* Reject unknown imported cities from queries
This commit fixes a bug where the city report returned `N/A` entries.
The functions that build imported data queries were using SQL
`COALESCE`, assuming city data is `NULL` when unknown, when actually its
unknown value is `0`.
This commit addresses the problem using SQL `NULLIF` combined with the
previous `COALESCE` call. With this change both `0` and `NULL` are
treated as unknown.
Since 1cb07efe6d cities can be `NULL`, but
previously we saved `0` as unknown.
Closes#1960
* Add entry to CHANGELOG
* Ignore cyclomatic complexity Credo check
This commit adds city data to imported records from Google Analytics. The
current implementation sets city to 0 because GA does not use the GeoNames
database.
Google Analytics Reporting API uses [Geographical IDs](https://developers.google.com/analytics/devguides/collection/protocol/v1/geoid)
to identify cities and countries. Plausible uses
[GeoNames](https://geonames.org/) and I couldn't find databases corelating the
two.
Fortunately, GA also returns the city name and this commit uses the city name
and the country ISO code to find the Geoname ID. To avoid making expensive ETS
searches, I created another ETS table in the Location library that uses
{country, city} as a key.
Related PR: https://github.com/plausible/location/pull/3
* Configure ingest repo access/pool size
If I'm not mistaken 3 is a sane default, the only
inserts we're doing are:
- session buffer dump
- events buffer dump
- GA import dump
And all are serializable within their scopes?
* Add IngestRepo
* Start IngestRepo
* Use IngestRepo for inserts
* Annotate ClickhouseRepo as read_only
So no insert* functions are expanded
* Update moduledoc
* rename alias
* Fix default env var value so it can be casted
* Use IngestRepo for migrations
* Set default ingest pool size from 3 to 5
in case conns are restarting or else...
* Ensure all Repo prometheus metrics are collected
* Make Device section components aware of (not set)
So that no extra sub-filters are possible when the unset
top item is selected.
* Support '(not set)' in breakdown/filters
* Update expectations for export tests
* Add extra tests for returning/filtering by '(not set)'
* Add changelog entry
* Remove ListReport conditional render
* Prevent redundant sub-filters
* Fix filter text rendering
---------
Co-authored-by: Uku Taht <uku.taht@gmail.com>
* Move Endpoint errors setup to common config
* Implement naive Sentry link resolver
* Implement error report e-mail
* Delete static sentry script
* Implement user feedback form on server errors
* Re-arrange pipe
* Use Sentry.Config.dsn() where applicable
* Fix typo
* Use Map.replace/3
* fix subquery for sessions in base_event_query/2
As the 'sessions' table is using the CollapsingMergeTree engine, we have
to select session_id's distinctively. Otherwise we will get multiple rows
(with sign -1 and 1) as long as the background merge hasn't happened.
* update changelog
* use GROUP BY instead of SELECT DISTINCT
* remove comma
* Fingerprint DBConnection.ConnectionError in Sentry
* Check events before creating a site
* Enable sites limit screen
* Remove debugging remnant
* Fix buggy assertions
This wasn't doing what expected:
iex(1)> Repo.exists?(Plausible.Site, domain: "foo")
[debug] QUERY OK source="sites" db=0.6ms idle=1906.2ms
SELECT TRUE FROM "sites" AS s0 LIMIT 1 []
* Encapsulate check to satisfy credo
* Use less technically involved error message
* Bring back e-mail to the limit error message
This commit stops logging `{:invalid_address, "ip"}` errors. This
reduces noise on Sentry, that is capturing error logs. Other errors are
still logged.
This PR replaces geolix with locus to simplify self-hosted setup. locus can auto-update maxmind dbs which are recommended for self-hosters if they want city-level geolocation. locus is also a bit faster.
This PR also uses a test mmdb file from https://github.com/maxmind/MaxMind-DB for e2e geolocation tests without stubs.
* Consolidate task timeouts, incrase to 15s
* Use Task.async_stream for parallel Clickhouse queries
* Propagate Opentelemetry context to child Task process
* Fix breakdown API pagination when using event metrics
This commit fixes a bug where the subsequent breakdown API pages had
the same items as the first page. The fix sorts the underlying
ClickHouse query by timestamp, keeping the same order between requests,
as we use OFFSET/LIMIT pagination.
* Fix repeated results assertion
* Add different ORDER BY to each breakdown property
* Ignore XX and T1 countries
* Add fallback if country_code=nil
* Lookup city overrides directly in CityOverrides module
* Changelog
* Add empty moduledoc
* Remove redundant comment
* Return empty list when breaking down by event:page without events
This commit fixes a bug with pagination where breaking down by event:page
would always return results despite pagination.
Closes#2255
* Update CHANGELOG.md
### Changes
Hello friends, I'm trying to self-host plausible on my server and have
observed that if I add a file named (for example) `secrets/LISTEN_IP` to
configure the bind IP I'll get the following error:
```
Compiling 1 file (.ex)
** (RuntimeError) Invalid LISTEN_IP '172.19.130.83
' error: :einval
/home/serafeim/plausible/config/runtime.exs:20: (file)
```
(notice that I get the same error when configuring the DATABASE_URL or
whatever comes first)
It seems that for whatever reason the File.read! will insert a newline
at the end of the file. Notice that I have double checked the file with
vim and *it does not contain a newline at the end* or at least I don't
know *how* to make the file not have a newline at the end.
Thus I'm adding a `String.trim()` there to fix that thing; even if I am
doing something wrong and my files *do* contain a newline at the end,
adding a `String.trim()` there definite won't hurt :)
Below you'll find a checklist. For each item on the list, check one
option and delete the other.
### Tests
- [ ] Automated tests have been added
- [x] This PR does not require tests
### Changelog
- [ ] Entry has been added to changelog
- [x] This PR does not make a user-facing change
### Documentation
- [ ] [Docs](https://github.com/plausible/docs) have been updated
- [x] This change does not need a documentation update
### Dark mode
- [ ] The UI has been tested both in dark and light mode
- [x] This PR does not change the UI
### Changes
This PR:
- pushes PromEx to the bottom of supervision stack to avoid Endpoint
instrumentation failure
- ensures the site cache is ready by exposing it through the health
check endpoint
- fixes event timestamps being calculated at compile time, with
regression unit and integration tests
### Tests
- [x] Automated tests have been added
- [ ] This PR does not require tests
### Changelog
- [ ] Entry has been added to changelog
- [x] This PR does not make a user-facing change
### Documentation
- [ ] [Docs](https://github.com/plausible/docs) have been updated
- [x] This change does not need a documentation update
### Dark mode
- [ ] The UI has been tested both in dark and light mode
- [x] This PR does not change the UI
* Update Sites.Cache
So it's now capable of refreshing most recent sites.
Refreshing a single site is no longer wanted.
* Introduce Warmer.RecentlyUpdated
This is Sites Cache warmer that runs only for
most recently updated sites every 30s.
* Validate Request creation early
* Rename RateLimiter to GateKeeper and introduce detailed policies
* Update events API tests - a provisioned site is now required
* Update events ingestion tests
* Make limits visible in CRM Sites index
* Hard-deprecate DOMAIN_BLACKLIST
* Remove unnecessary clause
* Fix typo
* Explicitly delegate Warmer.All
* GateKeeper.allwoance => GateKeeper.check
* Instrument Sites.Cache measurments
* Update send_pageview task to output response headers
* Instrument ingestion pipeline
* Credo
* Make event telemetry test a sync case
* Simplify Request.uri/hostname handling
* Use embedded schema, apply action and rely on get_field
* Parse event URL in Plausible.Ingestion.Request
* Parse event domain in Plausible.Ingestion.Request
* Rework ingestion pipeline processing (#2462)
* Rework ingestion pipeline processing
So that Request can have multiple domains and
based on that each event is processed uniformly.
The build_and_buffer/1 function now returns an
accumulator with all the dropped/buffered events
for further inspection.
* Reduce function complexity
* Don't chain struct fields to check for an empty host
* Separate referrer and utm tags
* Fix up `with` clause, credo was right cc @vinibrsl
Co-authored-by: Adam Rutkowski <hq@mtod.org>
* Allow refreshing a single site cache + clear cache on prefill
* Reorganize Site.Cache tests with describe blocks
* Tidy up Cache tests
* Make sure the cache is cleared on (p)re-fill
* Allow process name customization in Cache.Warmer
* s/Cache.prefill/Cache.refresh
* Unify Cache refresh instrumentation
* Apply credo suggestion: change `with` to `case`
* Update typespecs to pass dialyzer
* Implement sites by domain caching interface + warmer
* Add test
* Implement hit rate interface
* Add moduledocs
* Fix up typespec
* s/warmer/warmer_fn
* Extract measure_duration/2
* Fix up typespec
* Log errors and return nil on cache internal errors
* Fix up non-existing cache test
* Retrieve specific db columns when pre-filling the cache
* Reduce the subset of fields retrieved from the DB
See 63f3c6233d (r89871536)
* Move schemas to paths corresponding to their namespace
* Include rate-limiting site schema fields
* Expose CRM RW access to site rate-limiting fields
* Remove 👾
* Add basic moduledoc placeholders
This commit adds OTEL tracing to external HTTP requests made with
the HttpClient module. It records the elapsed time, host, path, and
response status code.
* Implement FF-driven DB lookup for sites during ingestion
We like to see the impact of doing a simple postgres lookup on each
ingestion event. The percentage-based feature flag `:ingestion_pg_lookup`
must be set in order for lookups to be executed.
* Fix resolving Cachex stats metrics
* Enable PromEx on dev env
This commit removes the current 1% sampling on ingestion to completely
ignore ingestion traces. We are not getting enough value from those, and
we'll keep dashboard, API and background jobs traces only.
People are likely to enter (copy/paste) goals from external sources
which can lead to whitespace characters appended by accident.
That renders the goal unusable and hard to distinct visually.
Normally to fix up existing goals we would use a data migration,
but this should be good enough to check if the problem
with never appearing goals resurfaces.
This commit fixes a bug where Google Analytics import tokens were not
being refreshed properly because the function was not returning the
expected tuple. Thanks to @aerosol we can nicely test this now.
This commit fixes a bug where users clearing one import and trying to
re-import would get invalid import date ranges. This was caused because
the stats_start_date field was not updated to nil after clearing
imported stats, as this field defines the end date of the import.
* Make TestUtils module available in all tests
* Add macros patching the application env in tests
Unfortunately a lot of existing functionality relies on
certain application env setup. This isn't ideal because
the app config is a shared state that prevents us from
running the tests in parallel.
Those macros encapsulate setting up new env for test purposes
and make sure the changes are reverted when the test finishes.
* Allow passing request opts to HTTPClient.post/4
We need this to swap custom request building in
Google Analytics import.
* Unify errors when listing sites
* React: propagate backend error messages if available
* React: catch API errors in Search Terms component
* Propagate google API errors on referrer drilldown
* Handle verified properties errors in SC settings
* Add missing tests for SC settings controller
* Unify errors for fetching search analytics queries (list stats)
* Unify errors refreshing Google Auth Token
* Test fetch_stats/3 errors and replace Double with Mox
* Fixup makrup
* s/class/className
* Simplify Search Terms display in case of errors
* Fix warnings