Prometheus Metrics Hybrid Chat
Some Hybrid Chat components expose metrics for Prometheus. These metrics can be plotted using Grafana or any other dashboard that supports Prometheus data source.
The details of currently available Prometheus Metrics are below:
Channel Manager
The Channel Manager 3.x exposes the following metrics for Prometheus:
| Metric | Request Direction | URL | Name | Tag Filter | Calculation Formula | Alarming | Description | |
|---|---|---|---|---|---|---|---|---|
| 1 | Total Inbound Requests Count per Channel | INBOUND | /ccm/actuator/prometheus | http_server_requests_seconds_count | Filter using uri tag to detect channel | The total number of messages received by the Channel Manager for every configured channel since the last restart. See URI mapping in the below table for filtering the channel. | ||
| 2 | Failed Inbound Requests Count | INBOUND | /ccm/actuator/prometheus | http_server_requests_seconds_count | Filter using uri tag to detect channelAND Filter using status != 200 tag to detect failed requests | The total number of messages for every configured channel that the Channel Manager failed to process since the last restart. See URI mapping in the below table for filtering the channel. | ||
| 3 | Successfully Processed Inbound Requests Count | INBOUND | /ccm/actuator/prometheus | http_server_requests_seconds_count | Filter using uri tag to detect channelAND Filter using status = 200 tag to detect failed requests | The total number of messages for every configured channel that the Channel manager successfully processed since the last restart. See URI mapping in the below table for filtering the channel. | ||
| 4 | Average Process Duration for Inbound Requests | INBOUND | /ccm/actuator/prometheus |
AND
| Filter using uri tag to detect channel | rate(http_server_requests_seconds_sum[5s])/ rate(http_server_requests_seconds_count[5s]) | ≥ 5 SECONDS | Average process duration that the Channel Manager took to process receiving inbound requests since last 5 seconds. See URI mapping in the below table for filtering the channel. |
| 5 | Total Outbound Messages Count per Channel | OUTBOUND | /ccm/actuator/prometheus | {channel}_outbound_request_total | Filter using {channel} name | The number of outbound messages processed by the Channel Manager since the last restart. See { channel} mapping in the below table for filtering the channel. | ||
| 6 | Successfully Processed Outbound Requests Count | OUTBOUND | /ccm/actuator/prometheus | {channel}_outbound_successful_request_total | Filter using {channel} name | All messages delivered by the Channel Manager to the specified channel since the last restart. If and when the Channel Manager cannot send a message, it will either retry (increasing failure retries) or will mark the message as invalid (increasing invalid messages count). See { channel} mapping in the below table for filtering the channel. | ||
| 7 | Total Invalid Outbound Messages Count | OUTBOUND | /ccm/actuator/prometheus | {channel}_outbound_invalid_request_total | Filter using {channel} name | Bad messages (bad request) received from the Chat Server that the 3rd party channel denied acceptance since the last restart. See { channel} mapping in the below table for filtering the channel. | ||
| 8 | Total Failed Outbound Messages Retries Count | OUTBOUND | /ccm/actuator/prometheus | {channel}_outbound_failed_request_total | Filter using {channel} name | ≥ 5 | Total number of retries that the Channel Manager took o process the last outbound message. See { channel} mapping in the below table for filtering the channel. | |
| 9 | Average Process Duration of Outbound Messages | OUTBOUND | /ccm/actuator/prometheus | {channel}_outbound_request_seconds_sumAND {channel}_outbound_request_seconds_count | Filter using {channel} name | rate({channel}_outbound_request_seconds_sum[5s])/ rate({channel}_outbound_request_seconds_count[5s]) | ≥ 5 SECONDS | The average process duration that the Channel Manager took to process an outbound message for the last 5 seconds. See { channel} mapping in the below table for filtering the channel. |
URI Mapping
For filtering incoming requests' Prometheus metrics, refer to the below URI mapping table
| Channel | Vendor | URI | Description | |
|---|---|---|---|---|
| 1 | Twilio | /msg-client/receive | ||
| 2 | 360Dilaog | /d360 | ||
| 3 | SMS/MMS | Twilio | /msg-client/receive | |
| 4 | Facebook Messenger | Facebook Graph API | /webhook | |
| 5 | Viber | Viber Rest API | /viber |
Channel-based Filtering
The following table shows the channel mapping with keywords that are used in the {channel} parameter in Prometheus outbound stats exposed by the Channel Manager.
| Channel | Vendor | Filter Keyword | Description | |
|---|---|---|---|---|
| 1 | Twilio | whatsapp_twilio | Use this keyword to filter outbound metrics per channel. e.g. whatsapp_twilio_outbound_requests_total provides count of total outbound requests for WhatsApp channel via Twilio. | |
| 2 | 360Dialog | whatsapp_d360 | Use this keyword to filter outbound metrics per channel. e.g. whatsapp_d360_outbound_requests_total provides count of total outbound requests for WhatsApp channel via Dialog360. | |
| 3 | SMS/MMS | Twilio | sms_twilio | Use this keyword to filter outbound metrics per channel. e.g. sms_twilio_outbound_requests_total provides count of total outbound requests for SMS/MMS channel via Twilio. |
| 4 | SMS | SMPP | smpp | Use this keyword to filter outbound metrics per channel. e.g. smpp_outbound_requests_total provides count of total outbound requests for SMS channel via an SMPP gateway. |
| 5 | Facebook Messenger | Facebook Graph API | facebook | Use this keyword to filter outbound metrics per channel. e.g. |
| 6 | Social Media Feed | Facebook Graph API | sm_feed | Use this keyword to filter outbound metrics per channel. e.g. sm_feed_outbound_requests_total provides count of total outbound requests for the Social Media Feed channel. |
| 7 | Viber | Viber Rest API | viber | Use this keyword to filter outbound metrics per channel. e.g. viber_outbound_requests_total provides count of total outbound requests for the Viber channel. |
Reporting Connector
The Reporting Connector 3.x exposes the following metrics for Prometheus:
| Name | URL | Description |
|---|---|---|
failed_query_count | <ip-address>:8496/api/metrics | The total number of failed queries in error.sql file. |
Chat Server
The Chat Server 3.x exposes the following metrics for Prometheus:
| Name | URL | Description |
|---|---|---|
conversation_count | /chat/metrics | Shows the currently active conversations, including the conversations waiting in queues. |
The metrics provided in this document are custom-based. For java and node.js technology-based metrics visit the relevant URL.
For database connector technology based metrics visit https://<fqdn>/database-connector/metric
For bot connector technology based metrics visit https://<fqdn>/database-connector/metrics