System status
How to read the “Status” page: version, services, background tasks, GEO databases, queues and integrations — and what to do about a warning.
The System status («Состояние системы») page shows whether the parts of the tracker are working: the databases, background tasks, click intake and event delivery. Come here when clicks do not appear in statistics, postbacks are not sent or the panel responds with errors. The section is visible to the owner and the admin.
How to open it and how the data refreshes
- Click the Admin panel («Админ-панель») button in the top right corner of the panel.
- Choose Status («Состояние»).
The page address is https://panel.example.com/maintenance. From the Settings («Настройки»), Team («Команда») and Logs («Логи») pages, the Status tab in the bar above the section leads here.
The upper part of the page refreshes itself every 15 seconds, and the Applying traffic settings («Применение настроек трафика») block and the queue blocks every 10 seconds. The Refresh («Обновить») button re-reads the upper part immediately.
Services and background tasks apply to the whole installation. The “CSV” line and the Event delivery («Доставка событий») table apply to the current workspace.
Color tells you where to look: green means all is well, yellow is a warning, red is a failure.
Version, uptime and memory
| Indicator | What it means |
|---|---|
| Version («Версия») | The panel build identifier, 12 characters. “Unknown” («Не определена») means the build did not pass an identifier |
| Admin uptime («Admin работает») | How many minutes have passed since the panel service last started. A small number when you did not restart anything means the service recently restarted on its own |
| Go heap | Memory used by the panel service, in MB. Next to the name is the Go version the service was built with |
| Checked («Проверено») | The time of the last check. If it does not change for longer than 15 seconds, the page is not getting a response |
The installed release number is shown in Settings → Updates («Настройки → Обновления»), see Updating the tracker from the panel.
Services and migrations
Three cards: postgres, clickhouse, redis.
| What the card shows | What it means |
|---|---|
| Available («Доступен») | The service responded to the check. Below are its version and data volume: “on disk” («на диске») for postgres and clickhouse, “in memory” («в памяти») for redis |
| Unavailable («Недоступен») | The service did not respond within the check time |
| Migrations · N («Миграции · N») | An expandable list of applied database structure updates. postgres and clickhouse have it |
What Unavailable leads to:
postgres— the panel cannot read settings, campaigns and the delivery queue.clickhouse— clicks and conversions are not written to statistics, and reports are not built. Accepted clicks wait in the intake queue.redis— there is no information from the click handler: the GEO database and click intake blocks will show a warning.
What to do: check the containers on the server with the commands from the Where to see the installation status section.
Background tasks
A table with the Task («Задача»), State («Состояние») and Last response («Последний отклик») columns.
| Task | What it does |
|---|---|
| Automation rules («Автоматические правила») | Checks the automation rules and performs their actions |
| CSV exports («Выгрузки CSV») | Prepares background report exports |
| Postback delivery («Доставка постбеков») | Sends outgoing S2S postbacks and Facebook events, records conversions |
| Target monitoring («Мониторинг целей») | Checks campaign offers and landing pages, see Campaign target diagnostics and monitoring |
| Audit log cleanup («Очистка журнала аудита») | Performs the daily cleanup, see Data retention |
| Domain check («Проверка доменов») | Checks the DNS and HTTPS of domains, see Domain check |
| Notifications («Уведомления») | Creates panel notifications about the license, backups, updates, offer caps, page checks, delivery queues and background services |
| Status | What it means | What to do |
|---|---|---|
| Running («Работает») | The task checked in recently without an error | Nothing |
| No response yet («Ещё нет отклика») | The task has never checked in since installation | Wait a few minutes and click Refresh |
| No recent response («Нет свежего отклика») | The last response is older than 3 minutes | Look at the admin container log on the server |
| Error («Ошибка») | The task's last run ended with an error | Look at the admin container log. The status returns to Running after a successful run |
| Waiting for license («Ожидает лицензию») | The Automation rules task checked in while the license was inactive: during that time rule actions are not performed. The panel is closed while the license is inactive, so this status is visible only right after renewal | Click Refresh: after the task's next run, the status changes to Running |
| Unknown status («Неизвестное состояние») | The task passed a value the panel does not know | Refresh the page |
The message “No confirmed workers telemetry” («Нет подтверждённой телеметрии workers») instead of the table means the panel could not read the task check-ins. Check the postgres card.
Under the table there is a “CSV” line with the number of the workspace's exports by state: pending — waiting, running — being prepared, ready — ready, failed — failed. “queue is empty” («очередь пуста») means there are no active exports. “state unknown” («состояние неизвестно») means the counter could not be read.
Engine GEO databases
Three columns: mmdb, ip2location, ip2proxy. Each has the status Loaded («Загружена»), Open error («Ошибка открытия») or Not connected («Не подключена»); connected ones also show the file size and date. The statuses and file replacement are explained on the Tracker GEO databases page.
The status shows which file the click handler actually opened. The file date does not confirm that the provider's data is fresh.
The line “No fresh data from redirect” («Нет свежих данных от redirect») instead of the columns means the information from the click handler is older than 30 seconds. Check the redirect container and the redis card.
Applying traffic settings
The block shows whether changes to campaigns, streams and domains have reached the click handler. The Applied («Применены») status means all is well. The other statuses and the server table are explained on the Traffic processing servers page.
Click intake queue
The Click intake («Приём кликов») table applies to the whole server. It has two queues: Clicks («Клики») and Landing page clicks («Переходы с лендинга»). The handler first saves a click to the server disk, then writes it to statistics in batches.
| Column | What it means |
|---|---|
| Batches («Пакеты») | How many batches are waiting to be written. A batch is not a single click |
| Oldest («Самый старый») | How long the oldest batch has been waiting |
| Journal / limit («Журнал / предел») | The volume of unwritten data on disk and the limit up to which the handler accepts new data |
| Free disk space («Свободно на диске») | Free space on the disk with the queue. “Unknown” («Неизвестно») means it could not be determined |
| Delivery errors («Ошибки доставки») | The number of failed writes to statistics since the handler started |
| Status | What it means | What to do |
|---|---|---|
| Running | Clicks are written without delay | Nothing |
| Delayed («Задержка») | The oldest batch has been waiting for longer than a minute | Check the clickhouse card. Clicks will appear in reports once they are written |
| Retrying after error («Повтор после ошибки») | Writing to statistics failed, and the handler is retrying | Check the clickhouse card |
| Journal almost at the intake limit («Журнал заполнен почти до предела приёма») | The journal has taken 80% of the intake limit or more | Restore writing urgently: check clickhouse and disk space |
| Low disk space («Мало места на диске») | Less than 1 GB is free | Free up space, see Backup retention and cleanup, disk space |
| Journal read error («Ошибка чтения журнала») | The handler could not read the queue on disk | Look at the redirect container log |
| Stale («Устарело») | The information from the handler is older than 30 seconds | Check the redirect container and the redis card |
In the Stale status, the line “No fresh signal from redirect. Queue state unknown.” («Нет свежего сигнала от redirect. Состояние очереди неизвестно.») is shown above the table. If there is no information from the handler at all, only this line remains, without the table. At the bottom of the block there are “Checked” and “Redirect signal” («Сигнал redirect») with the time of the latest information or the word “none” («нет»).
The Journal almost at the intake limit and Low disk space statuses mean a risk of losing traffic: when the queue on disk reaches the limit, the handler stops accepting new clicks, and the visitor gets a 503 error.
Event delivery
The table counts the unfinished tasks of the current workspace. The rows are task types: conversion — recording a conversion, s2s — an outgoing postback, capi — a Facebook event. Sent tasks are not included in the table. If there are no unfinished tasks, the line “No unfinished tasks” («Незавершённых задач нет») is shown.
| Column | What it means | What to do |
|---|---|---|
| Pending («Ожидают») | Tasks waiting to be sent or retried | Nothing, if the number is not growing. While the license is inactive, S2S and CAPI wait for renewal |
| In progress («В работе») | Tasks being sent right now | Nothing |
| Errors («Ошибки») | Tasks the tracker no longer retries on its own | Open Logs → Event delivery («Логи → Доставка событий»), see Manual retry |
| Processing timed out («Истёк срок обработки») | Tasks were taken into work but not completed in time | Check the Postback delivery task in the Background tasks («Фоновые задачи») block |
| Unknown result («Неизвестный результат») | The request was sent, but no response from the recipient was received | Before retrying, check the event on the recipient's side, see “Receipt not confirmed” («Получение не подтверждено») |
| Queue age («Возраст очереди») | How long the oldest unfinished task has been waiting | A growing value while the task is running is a reason to open the delivery log |
The same two tables are shown in Logs → Queue status («Логи → Состояние очередей»).
The “Integrations” block
The Integrations («Подключения») block has one item — Node management («Управление узлами»). Configured («Настроено») means the server has the key the tracker uses to encrypt access to reverse proxy nodes and the Cloudflare tokens. Not configured («Не настроено») means there is no key, and these sections do not work.
The item only shows whether the setting exists. Check the availability of the nodes themselves in the Reverse proxy nodes («Реверс-прокси ноды») section.
If the page did not load
A red line with the note “The data shown below may be outdated” («Показанные ниже данные могут быть устаревшими»)
The panel did not get a response from its service. An error text with this note appears above the blocks, and the blocks show the last values received. Click Refresh. If the error repeats, check the admin container on the server — the commands are in the Where to see the installation status section.
A red line above the queues with the note “The data below may be outdated” («Данные ниже могут быть устаревшими»)
The queue block did not load; the rest of the page works. The block retries the request on its own after 10 seconds. If the line does not go away, check the postgres card.
Couldn't check settings application («Не удалось проверить применение настроек»)
The Applying traffic settings block did not load. The request is repeated after 10 seconds. See the Traffic processing servers page for details.
The page is closed by a license message
While the license is inactive, the whole panel is closed, including Status. Traffic and conversion intake continue. See License expiry.