Activity Logs & Reports
Dashboard Activity Logs, task correlation, usage reports, and audit retention.
SmartGate records tool and API activity for security, debugging, and ROI reporting. Two Dashboard areas cover most observability needs:
| Area | Path | Best for |
|---|---|---|
| Activity Logs | Dashboard → Activity Logs | Per-request audit, Events vs Tasks views |
| Usage Reports | Dashboard → Usage Reports | Token trends, savings, module breakdown |
Pull-based audit (Dashboard + export) is the supported model today. Webhooks for push delivery are Preview only.
Activity Logs
Open Dashboard → Activity Logs.
Events vs Tasks
| Tab | Shows |
|---|---|
| Events | Individual API/MCP calls (compact or table density) |
| Tasks | Rows grouped by task correlation ID — multiple calls from one agent job |
Use Tasks when debugging multi-step agent runs (several fetches sharing one correlation).
ID glossary
| Concept | Header / field | Meaning |
|---|---|---|
| Request ID | Server-generated | One HTTP or MCP tool invocation |
| Task correlation | X-SmartGate-Correlation-Id | Groups spans in a single agent task |
| Session trace | X-SmartGate-Trace-Id | Optional; groups multiple tasks in a session |
Send correlation headers on REST calls to improve Task grouping. MCP calls are tagged automatically with route=mcp. Details: MCP Endpoint → Audit headers.
What gets logged
Typical fields include tool name, timestamp, route (mcp, rest, playground), success/failure, and optional metadata such as requested URL for fetch tools.
Access denials and authentication failures may also appear as security-relevant events.
Retention
Log retention depends on your plan (see Security):
| Plan | Retention |
|---|---|
| Free | 7 days |
| Pro | 30 days |
| Teams | 90 days |
| Enterprise | 180 days |
Older rows are purged automatically. Upgrade or export before retention expires for long-term archives.
Refresh and live tail
Use the toolbar Refresh control to load recent events. Live tail behavior follows the current Logs UI (auto-refresh where enabled).
Usage Reports
Open Dashboard → Usage Reports.
Time range
Select 7d, 30d, or 90d to change the reporting window.
Trends section
By default you see Token usage trend. Optional checkboxes can add other trends on the same date axis (for example cost savings or budget-related series — depending on your data).
This helps compare "tokens burned" vs "savings estimated" without switching pages.
Module comparison
A full-width chart breaks usage down by SmartGate module (fetch, search, compress, etc.).
Include Playground
Toggle Include Playground to add or exclude Playground REST tests from report totals. Preference may persist in your browser.
Relationship to Dashboard home
- Dashboard home — quick savings / quota snapshot
- Usage Reports — deeper trends and module mix
- Budget — monthly limits and enforcement settings
Compliance-oriented usage
For enterprise review:
- Require agents to send
X-SmartGate-Correlation-Idon multi-step REST workflows. - Review Tasks view weekly for anomalous fetch/search volume.
- Export or archive logs within your plan retention window.
- Do not rely on outbound webhooks for audit until delivery GA.
Contact support@smartgate.network for Enterprise export requirements.
Related
- Tools Reference — what each tool records
- Security — retention and rate limits
- Webhooks (Preview) — future push notifications
What an audit row represents
Each MCP or REST tool invocation can produce a team-scoped audit record: tool name, parameters (as stored), oken_used, latency, success/failure, and identifiers such as equest_id / eam_id. Activity Logs in the dashboard and /api/v1/audit/trace expose read paths; /api/v1/audit/export supports compliance workflows.
How to investigate a session
- Filter by time window and team.
- Isolate the API key or request id if you have it.
- Walk pipe steps in order (search, fetch, dedup, context_gate) when smart_pipe was used.
- Confirm compress rows expose savings fields when smart_context_gate ran.
- Export the window if a reviewer needs an artifact outside the retention UI.
Retention is a plan entitlement
Free, Pro, Teams, and Enterprise keep Activity Logs for different day counts from the plan catalog SSOT. Marketing pages and product UI read the same catalog. If you need longer retention, upgrade or export on a cadence (Security, Billing).
Metrics vs invoices
Usage Reports may show estimated gateway counterfactual savings (Est. USD saved) derived from audited tool metrics and an internal benchmark. Use that to compare tool paths and Gainshare inputs. Do not paste it into finance systems as a replacement for your model provider invoice.
Common pitfalls
- Looking at the wrong team filter after accepting an invite.
- Expecting Playground-only runs to match production MCP key traffic.
- Assuming deleted keys remove historical rows (rows remain for the retention window).
- Treating empty logs as gateway down when the host never completed MCP handshake (Connect).
Audit is the minimum bar for production agents: without it you cannot answer who fetched what, when, at what token cost.
Practical adoption notes
Treat SmartGate as shared infrastructure: one gateway URL, team-scoped keys, and audit on by default. Prefer pipe templates for repeatable research paths so Activity Logs show a coherent chain instead of ad-hoc tool spam. Keep host-model prompts focused on decisions; let Research and Context primitives prepare evidence. When FinOps asks for savings, export Usage Reports alongside the gateway benchmark explanation—never present Est. USD saved as a vendor LLM invoice. When security asks for evidence, export an audit window and show retention tied to the current plan catalog.
If you are migrating from scripts, replace raw HTTP fetch helpers first, then add dedup and compression, then budget guard. If you are migrating from a model gateway mindset, stop expecting chat completions on /api/mcp and wire MCP hosts instead. If you are writing human-facing product copy, lead with Smart Research / Smart Decision outcomes; keep MCP JSON on builder paths. These habits keep the site and the product aligned with the Positioning Bible: intelligence layer between AI and the world, not an LLM proxy.