/// <summary>
/// Constructor
/// </summary>
/// <param name="storageConnectionString">Storage account connection string</param>
/// <param name="azureTablePrefix">The repository stores data in two tables called httprequestbycorrelationid and httprequestbydatedescending, if this parameter is not null and not whitespace then it is used as a prefix for those table names.</param>
/// <param name="granularity">The level of granularity for data in the partition. On a low traffic site hourly or even daily can be useful, whereas busy sites minute or second are more useful.</param>
public AzureHttpLoggerRepository(string storageConnectionString, string azureTablePrefix, LogByDateGranularityEnum granularity)
{
if (string.IsNullOrWhiteSpace(storageConnectionString)) throw new ArgumentNullException(nameof(storageConnectionString));
if (azureTablePrefix == null)
{
azureTablePrefix = "";
}
CloudStorageAccount storageAccount = CloudStorageAccount.Parse(storageConnectionString);
CloudTableClient client = storageAccount.CreateCloudTableClient();
_byCorrelationIdTable = client.GetTableReference(string.Concat(azureTablePrefix, RequestByCorrelationIdTableName));
_byDateTimeDescendingTable = client.GetTableReference(string.Concat(azureTablePrefix, RequestByDateTimeDescendingTableName));
_byCorrelationIdTable.CreateIfNotExists();
_byDateTimeDescendingTable.CreateIfNotExists();
switch (granularity)
{
case LogByDateGranularityEnum.Hour:
_granularPartitionKeyFormat = "yyyy-MM-dd hh";
break;
case LogByDateGranularityEnum.Day:
_granularPartitionKeyFormat = "yyyy-MM-dd";
break;
case LogByDateGranularityEnum.Minute:
_granularPartitionKeyFormat = "yyyy-MM-dd hh:mm";
break;
case LogByDateGranularityEnum.Second:
_granularPartitionKeyFormat = "yyyy-MM-dd hh:mm:ss";
break;
}
}
/// <summary>
/// In order to capture the full set of data during a HTTP request this OWIN plugin must be the first plugin in the chain.
///
/// Configure OWIN to trace HTTP requests and responses. Note that capturing request query parameters and request and response data could cause
/// data security issues - potentially sensitive data will be logged to Azure table storage.
///
/// Therefore by default none of this is captured.
///
/// The HTTP traces are stored in these tables:
///
/// httprequestbydatedescending
/// httprequestbycorrelationid
///
/// If this causes you any conflicts then the table names can be given an optional prefix using the azurePrefix parameter.
///
/// Request and response data is stored in blob containers named httprequestdata and httpresponsedata respectively and the azurePrefix
/// parameter is also applied to these container names. Blob names are linked to the log item ID held within the tables.
/// </summary>
/// <param name="appBuilder">The app builder extended</param>
/// <param name="storageConnectionString">The storage string for </param>
/// <param name="captureRequestParams">True if you wish to capture query parameters, false if not.</param>
/// <param name="captureRequestData">True if you wish to capture request data, false if not.</param>
/// <param name="captureResponseData">True if you wish to capture response data, falise if not.</param>
/// <param name="captureRequestHeaders">To capture all request headers set a single array element of "*" otherwise specify the headers you wish to capture.</param>
/// <param name="captureResponseHeaders">To capture all response headers set a single array element of "*" otherwise specify the headers you wish to capture.</param>
/// <param name="httpCorrelationHeaderKey">
/// It can be helpful when calling across http boundaries to be able to tie together the flow of events with a correlation ID and by default the logger
/// looks for a correlation ID in the header correlation-id. If the header is missing then no correlation ID is used but the logger will work. If you
/// wish to disable this behaviour then set this to null or if you want to use a different header then set the header name here.
///
/// The HttpCorrelator middleware also in this assembly can be used to add a correlation ID if none is present and should be placed before the logger middleware
/// in the pipeline.
/// </param>
/// <param name="azurePrefix">If you need to avoid table name conflicts</param>
/// <param name="granularity">Sets the level of granularity on the partition key of the httprequestbydatedescending table. Busier your site likely the more granular you want
/// this to be. Defaults to hourly</param>
/// <returns></returns>
public static IAppBuilder UseAzureHttpLogger(this IAppBuilder appBuilder,
string storageConnectionString,
bool captureRequestParams = false,
bool captureRequestData = false,
bool captureResponseData = false,
string[] captureRequestHeaders = null,
string[] captureResponseHeaders = null,
string httpCorrelationHeaderKey = "correlation-id",
string azurePrefix = "",
LogByDateGranularityEnum granularity = LogByDateGranularityEnum.Hour)
{
appBuilder.Use(typeof(HttpLogger),
new AzureHttpLoggerRepository(storageConnectionString, azurePrefix, granularity),
captureRequestParams,
captureRequestData,
captureResponseData,
captureRequestHeaders,
captureResponseHeaders,
httpCorrelationHeaderKey);
return(appBuilder);
}
/// <summary>
/// In order to capture the full set of data during a HTTP request this OWIN plugin must be the first plugin in the chain.
///
/// Configure OWIN to trace HTTP requests and responses. Note that capturing request query parameters and request and response data could cause
/// data security issues - potentially sensitive data will be logged to Azure table storage.
///
/// Therefore by default none of this is captured.
///
/// The HTTP traces are stored in these tables:
///
/// httprequestbydatedescending
/// httprequestbycorrelationid
///
/// If this causes you any conflicts then the table names can be given an optional prefix using the azurePrefix parameter.
///
/// Request and response data is stored in blob containers named httprequestdata and httpresponsedata respectively and the azurePrefix
/// parameter is also applied to these container names. Blob names are linked to the log item ID held within the tables.
/// </summary>
/// <param name="appBuilder">The app builder extended</param>
/// <param name="storageConnectionString">The storage string for </param>
/// <param name="captureRequestParams">True if you wish to capture query parameters, false if not.</param>
/// <param name="captureRequestData">True if you wish to capture request data, false if not.</param>
/// <param name="captureResponseData">True if you wish to capture response data, falise if not.</param>
/// <param name="captureRequestHeaders">To capture all request headers set a single array element of "*" otherwise specify the headers you wish to capture.</param>
/// <param name="captureResponseHeaders">To capture all response headers set a single array element of "*" otherwise specify the headers you wish to capture.</param>
/// <param name="httpCorrelationHeaderKey">
/// It can be helpful when calling across http boundaries to be able to tie together the flow of events with a correlation ID and by default the logger
/// looks for a correlation ID in the header correlation-id. If the header is missing then no correlation ID is used but the logger will work. If you
/// wish to disable this behaviour then set this to null or if you want to use a different header then set the header name here.
///
/// The HttpCorrelator middleware also in this assembly can be used to add a correlation ID if none is present and should be placed before the logger middleware
/// in the pipeline.
/// </param>
/// <param name="azurePrefix">If you need to avoid table name conflicts</param>
/// <param name="granularity">Sets the level of granularity on the partition key of the httprequestbydatedescending table. Busier your site likely the more granular you want
/// this to be. Defaults to hourly</param>
/// <returns></returns>
public static IAppBuilder UseAzureHttpLogger(this IAppBuilder appBuilder,
string storageConnectionString,
bool captureRequestParams = false,
bool captureRequestData = false,
bool captureResponseData = false,
string[] captureRequestHeaders = null,
string[] captureResponseHeaders = null,
string httpCorrelationHeaderKey = "correlation-id",
string azurePrefix = "",
LogByDateGranularityEnum granularity = LogByDateGranularityEnum.Hour)
{
appBuilder.Use(typeof(HttpLogger),
new AzureHttpLoggerRepository(storageConnectionString, azurePrefix, granularity),
captureRequestParams,
captureRequestData,
captureResponseData,
captureRequestHeaders,
captureResponseHeaders,
httpCorrelationHeaderKey);
return appBuilder;
}
/// <summary>
/// Constructor
/// </summary>
/// <param name="storageConnectionString">Storage account connection string</param>
/// <param name="azureTablePrefix">The repository stores data in two tables called httprequestbycorrelationid and httprequestbydatedescending, if this parameter is not null and not whitespace then it is used as a prefix for those table names.</param>
/// <param name="granularity">The level of granularity for data in the partition. On a low traffic site hourly or even daily can be useful, whereas busy sites minute or second are more useful.</param>
public AzureHttpLoggerRepository(string storageConnectionString, string azureTablePrefix, LogByDateGranularityEnum granularity)
{
if (string.IsNullOrWhiteSpace(storageConnectionString))
{
throw new ArgumentNullException(nameof(storageConnectionString));
}
if (azureTablePrefix == null)
{
azureTablePrefix = "";
}
CloudStorageAccount storageAccount = CloudStorageAccount.Parse(storageConnectionString);
CloudTableClient client = storageAccount.CreateCloudTableClient();
_byCorrelationIdTable = client.GetTableReference(string.Concat(azureTablePrefix, RequestByCorrelationIdTableName));
_byDateTimeDescendingTable = client.GetTableReference(string.Concat(azureTablePrefix, RequestByDateTimeDescendingTableName));
_byCorrelationIdTable.CreateIfNotExists();
_byDateTimeDescendingTable.CreateIfNotExists();
switch (granularity)
{
case LogByDateGranularityEnum.Hour:
_granularPartitionKeyFormat = "yyyy-MM-dd hh";
break;
case LogByDateGranularityEnum.Day:
_granularPartitionKeyFormat = "yyyy-MM-dd";
break;
case LogByDateGranularityEnum.Minute:
_granularPartitionKeyFormat = "yyyy-MM-dd hh:mm";
break;
case LogByDateGranularityEnum.Second:
_granularPartitionKeyFormat = "yyyy-MM-dd hh:mm:ss";
break;
}
}
