Interface IApplicationInsightsService
- Namespace
- Quilt4Net.Toolkit.Features.ApplicationInsights
- Assembly
- Quilt4Net.Toolkit.dll
public interface IApplicationInsightsServiceMethods
CanConnectAsync(IApplicationInsightsContext)
Task<bool> CanConnectAsync(IApplicationInsightsContext context)Parameters
contextIApplicationInsightsContext
Returns
GetCapTimelineAsync(IApplicationInsightsContext, int, CancellationToken)
Daily ingestion-cap timeline over the last days days: the configured daily
cap (GB) plus, per day, the ingested volume, the cap-hit time (from the workspace
"data collection stopped due to daily limit" operation events), and an estimate of the
uncapped volume. Degrades to an empty timeline when the Operation/Usage tables aren't readable.
Task<CapTimeline> GetCapTimelineAsync(IApplicationInsightsContext context, int days, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContextdaysintcancellationTokenCancellationToken
Returns
GetClusterNodeCpuAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
Kubernetes node CPU-used percentage per node (k8s.node.name) over
timeSpan. Computed as 100 * k8s.node.cpu.usage / k8s.node.allocatable_cpu
(cores used / total schedulable cores; on k3s allocatable == physical cores). Only bins where
allocatable_cpu is present render, so coverage starts when the collector began emitting
it. Series is the node name. Bin size scales with the window.
IAsyncEnumerable<MetricSample> GetClusterNodeCpuAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetClusterNodeFilesystemAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
Kubernetes node filesystem-used percentage per node (k8s.node.name) over
timeSpan. Computed as 100 * usage / capacity from kubeletstats
k8s.node.filesystem.usage and k8s.node.filesystem.capacity — answers "is a
node running out of disk". Series is the node name. Bin size
scales with the window. (Host disk I/O is intentionally not surfaced: the cluster Linux
nodes' system.disk.* rows carry no host identity, so it cannot be attributed.)
IAsyncEnumerable<MetricSample> GetClusterNodeFilesystemAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetClusterNodeMemoryAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
Kubernetes node memory-used percentage per node (k8s.node.name) over
timeSpan. Computed per bin as
100 * usage / (usage + available) from kubeletstats k8s.node.memory.usage
and k8s.node.memory.available (there is no direct capacity metric; the pair sums to
total). Series is the node name. Bin size scales with the window.
IAsyncEnumerable<MetricSample> GetClusterNodeMemoryAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetClusterPodCpuAsync(IApplicationInsightsContext, string, TimeSpan, CancellationToken)
Per-pod CPU usage in cores for the pods scheduled on node
(k8s.node.name) over timeSpan. Source: kubeletstats
k8s.pod.cpu.usage. This is the node→pod drill-down for
GetClusterNodeCpuAsync(IApplicationInsightsContext, TimeSpan, CancellationToken). Series is
{k8s.namespace.name}/{k8s.pod.name}. Bin size scales with the window.
IAsyncEnumerable<MetricSample> GetClusterPodCpuAsync(IApplicationInsightsContext context, string node, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContextnodestringtimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetClusterPodMemoryAsync(IApplicationInsightsContext, string, TimeSpan, CancellationToken)
Per-pod memory usage in MB for the pods scheduled on node
(k8s.node.name) over timeSpan. Source: kubeletstats
k8s.pod.memory.usage (absolute bytes — pods have no consistent capacity to render a
percentage against, unlike nodes). This is the node→pod drill-down for
GetClusterNodeMemoryAsync(IApplicationInsightsContext, TimeSpan, CancellationToken). Series is
{k8s.namespace.name}/{k8s.pod.name}. Bin size scales with the window.
IAsyncEnumerable<MetricSample> GetClusterPodMemoryAsync(IApplicationInsightsContext context, string node, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContextnodestringtimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetClusterTotalCpuAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
Whole-cluster CPU-used percentage over timeSpan, capacity-weighted:
100 * Σ usage / Σ allocatable_cpu across all nodes per bin (a 12-core node contributes
proportionally more than a 2-core node). Source: k8s.node.cpu.usage /
k8s.node.allocatable_cpu. Series is "Cluster".
IAsyncEnumerable<MetricSample> GetClusterTotalCpuAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetClusterTotalFilesystemAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
Whole-cluster total filesystem-used percentage over timeSpan,
weighted by capacity: 100 * Σ usage / Σ capacity across all nodes per bin. Source:
k8s.node.filesystem.usage / k8s.node.filesystem.capacity.
Series is "Cluster".
IAsyncEnumerable<MetricSample> GetClusterTotalFilesystemAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetClusterTotalMemoryAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
Whole-cluster total memory-used percentage over timeSpan, weighted
by capacity: 100 * Σ usage / Σ (usage + available) across all nodes per bin (so larger
nodes dominate the total). Source: k8s.node.memory.usage / k8s.node.memory.available.
Series is "Cluster".
IAsyncEnumerable<MetricSample> GetClusterTotalMemoryAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetCountAsync(IApplicationInsightsContext, string, TimeSpan, CancellationToken)
IAsyncEnumerable<CountData> GetCountAsync(IApplicationInsightsContext context, string environment, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContextenvironmentstringtimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetCpuUtilizationAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
CPU-busy percentage per host (cloud_RoleInstance) over timeSpan.
Derived from the OpenTelemetry system.cpu.utilization idle reading:
100 * (1 - avg(idle)). Bin size scales with the window (see Quilt4Net.Toolkit.Features.ApplicationInsights.MetricsBinSelector).
IAsyncEnumerable<MetricSample> GetCpuUtilizationAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetDetail(IApplicationInsightsContext, string, LogSource, string, TimeSpan)
Task<LogDetails> GetDetail(IApplicationInsightsContext context, string id, LogSource source, string environment, TimeSpan timeSpan)Parameters
contextIApplicationInsightsContextidstringsourceLogSourceenvironmentstringtimeSpanTimeSpan
Returns
GetDiskCapacityAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
Per-volume disk capacity (total GB, plus the free/used/reserved breakdown) across
timeSpan. Sourced from system.filesystem.usage by summing the
three states per volume — gives the UI a real "% of capacity" reference for free-space bars
instead of needing to scale them relative to whichever host happens to have the most space.
Series label matches GetDiskFreeAsync(IApplicationInsightsContext, TimeSpan, CancellationToken) so consumers can join the two by
Series / Series.
IAsyncEnumerable<DiskCapacity> GetDiskCapacityAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetDiskFreeAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
Filesystem free (GB) per host+device over timeSpan. Source: OpenTelemetry
system.filesystem.usage with state=free — operators want the "how much room is left"
view, where a line trending toward zero means the disk is filling up. Series label is
{host.name} {device} so multiple volumes on one host show as separate lines. Bin
size scales with the window.
IAsyncEnumerable<MetricSample> GetDiskFreeAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetEnvironments(IApplicationInsightsContext)
IAsyncEnumerable<EnvironmentOption> GetEnvironments(IApplicationInsightsContext context)Parameters
contextIApplicationInsightsContext
Returns
GetLogCountByServiceAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
Counts log entries per (service × severity × environment × source) across AppTraces,
AppExceptions and AppRequests over timeSpan. Severity is
unified across sources: traces use their own SeverityLevel; exceptions are always
counted as Error; requests use Information when Success is true
and Error otherwise. Service is coalesced from Properties.ApplicationName then
AppRoleName.
No server-side env / source filter — admin UIs typically want every dimension at hand so
flipping a filter checkbox can be done locally without an extra round trip. Pre-filtering
is one .Where away at the call site.
IAsyncEnumerable<LogCountByServiceCell> GetLogCountByServiceAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetMeasureAsync(IApplicationInsightsContext, string, TimeSpan, CancellationToken)
IAsyncEnumerable<MeasureData> GetMeasureAsync(IApplicationInsightsContext context, string environment, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContextenvironmentstringtimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetMemoryUtilizationAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
Memory-used percentage per host (cloud_RoleInstance) over timeSpan.
Source: OpenTelemetry system.memory.utilization with state=used. Bin size scales with
the window (see Quilt4Net.Toolkit.Features.ApplicationInsights.MetricsBinSelector).
IAsyncEnumerable<MetricSample> GetMemoryUtilizationAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetNetworkThroughputAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
Network throughput (MB/s) per host (cloud_RoleInstance) over timeSpan.
Computed as the per-bin delta of system.network.io divided by the bin duration; negative
deltas (host restart / counter reset) are dropped. Bin size scales with the window.
IAsyncEnumerable<MetricSample> GetNetworkThroughputAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetSummaries(IApplicationInsightsContext, string, TimeSpan, CancellationToken)
IAsyncEnumerable<SummarySubset> GetSummaries(IApplicationInsightsContext context, string environment, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContextenvironmentstringtimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetSummary(IApplicationInsightsContext, string, LogSource, string, TimeSpan, int, string)
Get the latest items for a single fingerprint. Capped at maxItems
(default 100) — a fingerprint can match 100k+ rows over the lookback window, and
returning all of them is what made detail/summary navigation hang. Items are ordered
newest-first; only the top maxItems are returned.
application scopes the instances to a single application (matching the
summary row); null/empty means no application filter.
Task<SummaryData> GetSummary(IApplicationInsightsContext context, string fingerprint, LogSource source, string environment, TimeSpan timeSpan, int maxItems = 100, string application = null)Parameters
contextIApplicationInsightsContextfingerprintstringsourceLogSourceenvironmentstringtimeSpanTimeSpanmaxItemsintapplicationstring
Returns
GetVersionMatrixAsync(IApplicationInsightsContext, TimeSpan?, bool)
Get the latest version of each (application, environment) combination
seen in the workspace. Tries the Quilt4NetStartup-tagged fast path
first, then falls back to a general scan for cells the fast path
did not cover. Pass null for lookback to
query the workspace's full retention.
IAsyncEnumerable<VersionMatrixCell> GetVersionMatrixAsync(IApplicationInsightsContext context, TimeSpan? lookback = null, bool forceRefresh = false)Parameters
contextIApplicationInsightsContextlookbackTimeSpan?forceRefreshbool
Returns
GetVolumeBySourceAsync(IApplicationInsightsContext, DateTimeOffset, DateTimeOffset, CancellationToken)
Billed ingestion volume (MB) per source table over an absolute UTC window — used for the
"Today" / "Yesterday" (UTC) ranges. Same Usage-table source and graceful-degrade behaviour
as the relative overload.
IAsyncEnumerable<VolumeBySource> GetVolumeBySourceAsync(IApplicationInsightsContext context, DateTimeOffset fromUtc, DateTimeOffset toUtc, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContextfromUtcDateTimeOffsettoUtcDateTimeOffsetcancellationTokenCancellationToken
Returns
GetVolumeBySourceAsync(IApplicationInsightsContext, TimeSpan, CancellationToken)
Billed ingestion volume (MB) per source table over timeSpan, from the
workspace Usage table — billing-grade and cheap (no telemetry scan). Backs the logging
cost dashboard's per-source breakdown. Returns empty when the Usage table can't be read
(permissions) rather than throwing, so the view degrades gracefully.
IAsyncEnumerable<VolumeBySource> GetVolumeBySourceAsync(IApplicationInsightsContext context, TimeSpan timeSpan, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContexttimeSpanTimeSpancancellationTokenCancellationToken
Returns
GetVolumeTimelineAsync(IApplicationInsightsContext, DateTimeOffset, DateTimeOffset, CancellationToken)
Billed ingestion volume (MB) per source table per UTC day over an absolute window — backs the
multi-day logging-volume line chart. Same Usage-table source and graceful-degrade behaviour.
IAsyncEnumerable<VolumeTimelinePoint> GetVolumeTimelineAsync(IApplicationInsightsContext context, DateTimeOffset fromUtc, DateTimeOffset toUtc, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContextfromUtcDateTimeOffsettoUtcDateTimeOffsetcancellationTokenCancellationToken
Returns
SearchAsync(IApplicationInsightsContext, string, string, TimeSpan, SeverityLevel, CancellationToken)
IAsyncEnumerable<LogItem> SearchAsync(IApplicationInsightsContext context, string environment, string text, TimeSpan timeSpan, SeverityLevel minSeverityLevel = SeverityLevel.Verbose, CancellationToken cancellationToken = default)Parameters
contextIApplicationInsightsContextenvironmentstringtextstringtimeSpanTimeSpanminSeverityLevelSeverityLevelcancellationTokenCancellationToken
Returns
SearchByCorrelationIdAsync(IApplicationInsightsContext, string, TimeSpan)
Returns all telemetry rows whose OperationId or
Properties.CorrelationId matches correlationId.
Searches AppTraces, AppExceptions, and AppRequests over the supplied
lookback window.
IAsyncEnumerable<LogItem> SearchByCorrelationIdAsync(IApplicationInsightsContext context, string correlationId, TimeSpan timeSpan)Parameters
contextIApplicationInsightsContextcorrelationIdstringtimeSpanTimeSpan
Returns
SearchByIncidentIdAsync(IApplicationInsightsContext, string, TimeSpan)
Returns all telemetry rows whose Properties.IncidentId matches
incidentId. Searches AppTraces, AppExceptions, and
AppRequests over the supplied lookback window. Used by the MCP
quilt4net.lookup_incident tool to map a user-facing incident id
(e.g. from a UI alert) back to the structured log entry that produced it.
IAsyncEnumerable<LogItem> SearchByIncidentIdAsync(IApplicationInsightsContext context, string incidentId, TimeSpan timeSpan)Parameters
contextIApplicationInsightsContextincidentIdstringtimeSpanTimeSpan