|
| 1 | +--- |
| 2 | +title: Delivery Overview |
| 3 | +--- |
| 4 | + |
| 5 | +Delivery Overview is a visual observability tool designed to help Segment users diagnose event delivery issues for any cloud-streaming destination receiving events from cloud-streaming sources. |
| 6 | + |
| 7 | +> info "Delivery Overview is currently in beta" |
| 8 | +> This means that the Delivery Overview feature is in active development, and some functionality may change before it becomes generally available. Beta users of Delivery Overview will still have access to the Event Delivery tab. During the public beta, Delivery Overview will only support event-streaming, cloud-mode destinations. |
| 9 | +> |
| 10 | +> The Delivery Overview public beta will be released to all Segment users on May 23, 2023 in a phased rollout. Users can expect to see the Delivery Overview tab in their destination details page by 5 PM PDT on May 26, 2023. |
| 11 | +
|
| 12 | +## Key features |
| 13 | + |
| 14 | +Delivery Overview has three core features: |
| 15 | +- [Pipeline view](#pipeline-view): a visual overview of each step your data takes during the delivery process |
| 16 | +- [Breakdown table](#breakdown-table): contains more detail about the events that were processed at each pipeline step |
| 17 | +- [Discard table](#discard-table): contains details about the events that failed or were filtered out of your process and allows you to inspect samples of them |
| 18 | + |
| 19 | +You can refine these tables using the time picker and the metric toggle, located under the destination header. With the time picker, you can specify a time period (last 10 minutes, 1 hour, 24 hours, or 7 days) for which you'd like to see data. With the metric toggle, you can switch between seeing metrics represented as percentages (for example, *85% of events* or *a 133% increase in events*) or as counts (*13 events* or *an increase of 145 events*.) Delivery Overview shows percentages by default. |
| 20 | + |
| 21 | +### Pipeline view |
| 22 | +The pipeline view provides insights into each step your data is processed by enroute to the destination, with an emphasis on the steps where data can be discarded due to errors or your filter preferences. Each step provides details into counts, change rates, and event details (like the associated Event Type or Event Names), and the discard steps (Failed on ingest, Filtered at source, Filtered at destination, & Failed delivery) provide you with the reasons events were dropped before reaching the destination. Discard steps also include how to control or alter that outcome, when possible. |
| 23 | + |
| 24 | +The pipeline view shows the following steps: |
| 25 | +- **Successfully received**: Events that Segment ingested from your source |
| 26 | +- **Failed on ingest**: Events that Segment received, but were dropped due to internal data validation rules |
| 27 | +- **Filtered at source**: Events that were discarded due to schema settings or [Protocols](/docs/protocols/) tracking plans |
| 28 | +- **Filtered at destination**: Events that were discarded due to [Destination Filters](/docs/guides/filtering-data/#destination-filters), [filtering in the Integrations object](/docs/guides/filtering-data/#filtering-with-the-integrations-object), or [per source schema integration filters](/docs/guides/filtering-data/#per-source-schema-integrations-filters). Actions destinations also have a filtering capability: for example, if your action is set to only send Identify events, all other event types will be filtered out. |
| 29 | +- **Failed delivery**: Events that have been discarded due to errors or unmet destination requirements |
| 30 | +- **Successful delivery**: Events that were successfully delivered to the destination |
| 31 | + |
| 32 | +### Breakdown table |
| 33 | +The breakdown table provides you with greater detail about the selected events. |
| 34 | + |
| 35 | +To open the breakdown table, select either the first step in the pipeline view (successfully received,) the last step in the pipeline view (successful delivery,) or select a discard step and then click on a discard reason. |
| 36 | + |
| 37 | +The breakdown table displays the following details: |
| 38 | +- **Event type**: The Segment spec event type (Track call vs. Identify call, for example) |
| 39 | +- **Event name**: The event name, provided by you or the source (*not available for inspection at all steps*) |
| 40 | +- **App version**: The app/release version, provided by you or the source (*not available for inspection at all steps*) |
| 41 | +- **Event count**: How many of each event either successfully made it through this pipeline step (in the case of the first or last steps in the pipeline view) or were filtered out (if you access it from a discard table) |
| 42 | +- **% Change**: Insight into how the event counts differ from the last comparable time range as a percentage<sup>1</sup> |
| 43 | + |
| 44 | +<sup>1:</sup> *Segment calculates the related change percentage by subtracting the percent of events impacted in the previous time period from the percent of impacted events in the current time period. For example, if last week 15% of your events were filtered at a source, but this week, 22% of your events were filtered at a source, you would have a related change percentage of 7%.* |
| 45 | + |
| 46 | +### Discard table |
| 47 | +The discard table provides you with greater detail about the events that failed to deliver or were filtered out of your sources and destinations. |
| 48 | + |
| 49 | +To open the discard table, click on one of the discard steps. If you click on a row in the discard table, you can see the breakdown table for the discarded events. |
| 50 | + |
| 51 | +The discard table displays the following details: |
| 52 | +- **Discard reason**: Any relevant error code, message, or description associated with the event's failure. When possible, Delivery Overview will link to any troubleshooting information you can use to get your events up and running again. Clicking on a discard reason brings you to the [breakdown table](#breakdown-table,) where you can see more detail about discarded events. |
| 53 | +- **Details & Samples**: View up to ten samples over the selected time range. Examine the error message and reason for the error or discard and inspect the payloads involved with the attempted transaction (*not available for inspection at all steps*) |
| 54 | +- **Event count**: How many of each event were discarded in this pipeline step |
| 55 | +- **% Change**: Insight into how the event counts differ from the last comparable time range as a percentage<sup>1</sup> |
| 56 | + |
| 57 | +<sup>1:</sup> *Segment calculates the related change percentage by subtracting the percent of events impacted in the previous time period from the percent of impacted events in the current time period. For example, if last week 15% of your events were filtered at a source, but this week, 22% of your events were filtered at a source, you would have a related change percentage of 7%.* |
| 58 | + |
| 59 | +## When should I use Delivery Overview? |
| 60 | +Delivery Overview is useful to diagnose delivery errors in the following scenarios: |
| 61 | +- **When setting up a destination, tracking plan, or filter for the first time**: With Delivery Overview, you can verify that the data you're sending to a new destination, a new tracking plan, or a new filter arrives in your destination as expected. |
| 62 | +- **When data is missing from your destination**: The pipeline view can help you see where your data is getting "stuck" on the way to your destination, which can help you quickly diagnose and address problems in your data pipeline. |
| 63 | +- **When emission or delivery volume fluctuates out of expected norms**: Delivery Overview will highlight where the largest rate change(s) occurred and what events were associated with the change. |
| 64 | + |
| 65 | +## Where do I find Delivery Overview? |
| 66 | +To view the Delivery Overview page: |
| 67 | +1. Sign into Segment. |
| 68 | +2. From the homepage, navigate to **Connection** > **Destinations** and click on the destination you'd like to investigate. |
| 69 | +3. Select the **Delivery Overview** tab from the destination header. |
| 70 | + |
| 71 | +## How do I use Delivery Overview? |
| 72 | +To use Delivery Overview: |
| 73 | + |
| 74 | +1. Navigate to the destination you'd like to review, and select **Delivery Overview** from the destination header. |
| 75 | +2. On the **Delivery Overview** tab, select a time period from the time picker. <br/> ___Optional___: *Turn the metric toggle off if you'd like to see the quantity of events as counts instead of percentages. Delivery Overview shows percentages by default.* |
| 76 | +3. Select a success or discard step to view additional context about the events that passed through that step. |
| 77 | + |
| 78 | +## How does Delivery Overview differ from other Segment monitoring and observability products? |
| 79 | +With Source Debugger or Event Delivery, you can only verify that events are successfully making it from your source or to your destination. If events fail, you have to troubleshoot to see where in the pipeline your events are getting stuck. With Event Tester, you can verify that your event makes it from your source to your destination, but if the results aren't what you expected, you're stuck troubleshooting your source, filters, tracking plans, and destinations. |
| 80 | + |
| 81 | +With Delivery Overview, you can verify that your source receives your events, that any filters and tracking plans work as expected, and that events successfully make it to your destination. Any errors or unexpected behavior can be identified using the pipeline view, leading to quicker resolution. |
| 82 | + |
| 83 | +## How can I configure alerts? |
| 84 | +During the Delivery Overview beta, you can use the Event Delivery alerting features (Delivery Alerts) by selecting the **Alerts** tab in the destination header. |
| 85 | + |
0 commit comments