SERVER-55104 Report time-series bucket granularity

9 files changed, 73 insertions, 16 deletions

diff --git a/jstests/noPassthroughWithMongod/timeseries_create.js b/jstests/noPassthroughWithMongod/timeseries_create.js
index fabe63f87b9..6ff50d9df40 100644
--- a/jstests/noPassthroughWithMongod/timeseries_create.js
+++ b/jstests/noPassthroughWithMongod/timeseries_create.js
@@ -100,6 +100,9 @@ testValidTimeseriesOptions({timeField: "time", metaField: "meta"});
 testValidTimeseriesOptions({timeField: "time", expireAfterSeconds: NumberLong(100)});
 testValidTimeseriesOptions(
     {timeField: "time", metaField: "meta", expireAfterSeconds: NumberLong(100)});
+testValidTimeseriesOptions({timeField: "time", metaField: "meta", granularity: "seconds"});
+testValidTimeseriesOptions(
+    {timeField: "time", metaField: "meta", granularity: "seconds", bucketMaxSpanSeconds: 3600});
 
 testInvalidTimeseriesOptions("", ErrorCodes.TypeMismatch);
 testInvalidTimeseriesOptions({timeField: 100}, ErrorCodes.TypeMismatch);
@@ -114,6 +117,10 @@ testInvalidTimeseriesOptions({timeField: "time", invalidOption: {}}, 40415);
 testInvalidTimeseriesOptions({timeField: "sub.time"}, ErrorCodes.InvalidOptions);
 testInvalidTimeseriesOptions({timeField: "time", metaField: "sub.meta"}, ErrorCodes.InvalidOptions);
 testInvalidTimeseriesOptions({timeField: "time", metaField: "time"}, ErrorCodes.InvalidOptions);
+testInvalidTimeseriesOptions({timeField: "time", metaField: "meta", granularity: "minutes"},
+                             ErrorCodes.InvalidOptions);
+testInvalidTimeseriesOptions({timeField: "time", metaField: "meta", bucketMaxSpanSeconds: 10},
+                             ErrorCodes.InvalidOptions);
 
 testCompatibleCreateOptions({storageEngine: {}});
 testCompatibleCreateOptions({indexOptionDefaults: {}});
diff --git a/src/mongo/db/SConscript b/src/mongo/db/SConscript
index 00520a3b284..865b287cda0 100644
--- a/src/mongo/db/SConscript
+++ b/src/mongo/db/SConscript
@@ -1132,6 +1132,7 @@ env.Library(
         '$BUILD_DIR/mongo/db/commands/fsync_locked',
         '$BUILD_DIR/mongo/db/repl/tenant_migration_access_blocker',
         '$BUILD_DIR/mongo/idl/server_parameter',
+        'catalog/database_holder',
         'commands/server_status_core',
         'service_context',
         'write_ops',
diff --git a/src/mongo/db/catalog/create_collection.cpp b/src/mongo/db/catalog/create_collection.cpp
index 66a8a62e799..5df1ec26d68 100644
--- a/src/mongo/db/catalog/create_collection.cpp
+++ b/src/mongo/db/catalog/create_collection.cpp
@@ -123,6 +123,16 @@ Status _createTimeseries(OperationContext* opCtx,
 
     options.viewOn = bucketsNs.coll().toString();
 
+    auto granularity = options.timeseries->getGranularity();
+    uassert(ErrorCodes::InvalidOptions,
+            "Time-series 'granularity' is required to be 'seconds'",
+            granularity == BucketGranularityEnum::Seconds);
+
+    auto bucketMaxSpan = options.timeseries->getBucketMaxSpanSeconds();
+    uassert(ErrorCodes::InvalidOptions,
+            "Time-series 'bucketMaxSpanSeconds' is required to be 3600",
+            bucketMaxSpan == 3600);
+
     if (options.timeseries->getMetaField()) {
         options.pipeline =
             BSON_ARRAY(BSON("$_internalUnpackBucket"
diff --git a/src/mongo/db/namespace_string.cpp b/src/mongo/db/namespace_string.cpp
index b8424d7361e..2414afacd6b 100644
--- a/src/mongo/db/namespace_string.cpp
+++ b/src/mongo/db/namespace_string.cpp
@@ -315,6 +315,11 @@ NamespaceString NamespaceString::makeTimeseriesBucketsNamespace() const {
     return {db(), kTimeseriesBucketsCollectionPrefix.toString() + coll()};
 }
 
+NamespaceString NamespaceString::bucketsNamespaceToTimeseries() const {
+    invariant(isTimeseriesBucketsCollection());
+    return {db(), coll().substr(kTimeseriesBucketsCollectionPrefix.size())};
+}
+
 bool NamespaceString::isReplicated() const {
     if (isLocal()) {
         return false;
diff --git a/src/mongo/db/namespace_string.h b/src/mongo/db/namespace_string.h
index fdcc41739c6..f918180f567 100644
--- a/src/mongo/db/namespace_string.h
+++ b/src/mongo/db/namespace_string.h
@@ -357,6 +357,11 @@ public:
     NamespaceString makeTimeseriesBucketsNamespace() const;
 
     /**
+     * Returns the time-series view namespace for this buckets namespace.
+     */
+    NamespaceString bucketsNamespaceToTimeseries() const;
+
+    /**
      * Returns whether a namespace is replicated, based only on its string value. One notable
      * omission is that map reduce `tmp.mr` collections may or may not be replicated. Callers must
      * decide how to handle that case separately.
diff --git a/src/mongo/db/timeseries/bucket_catalog.cpp b/src/mongo/db/timeseries/bucket_catalog.cpp
index 4f25cb9cc92..ed86ce6441f 100644
--- a/src/mongo/db/timeseries/bucket_catalog.cpp
+++ b/src/mongo/db/timeseries/bucket_catalog.cpp
@@ -138,13 +138,13 @@ StatusWith<std::shared_ptr<BucketCatalog::WriteBatch>> BucketCatalog::insert(
             return true;
         }
         auto bucketTime = (*bucket).getTime();
-        if (time - bucketTime >= kTimeseriesBucketMaxTimeRange) {
+        if (time - bucketTime >= Seconds(options.getBucketMaxSpanSeconds())) {
             stats->numBucketsClosedDueToTimeForward.fetchAndAddRelaxed(1);
             return true;
         }
         if (time < bucketTime) {
             if (!(*bucket)->hasBeenCommitted() &&
-                (*bucket)->latestTime - time < kTimeseriesBucketMaxTimeRange) {
+                (*bucket)->latestTime - time < Seconds(options.getBucketMaxSpanSeconds())) {
                 (*bucket).setTime();
             } else {
                 stats->numBucketsClosedDueToTimeBackward.fetchAndAddRelaxed(1);
diff --git a/src/mongo/db/timeseries/bucket_catalog.h b/src/mongo/db/timeseries/bucket_catalog.h
index a17d01f9d36..9f2b00befaa 100644
--- a/src/mongo/db/timeseries/bucket_catalog.h
+++ b/src/mongo/db/timeseries/bucket_catalog.h
@@ -47,10 +47,6 @@ class BucketCatalog {
     using IdleList = std::list<Bucket*>;
 
 public:
-    // This constant, together with parameters defined in timeseries.idl, defines limits on the
-    // measurements held in a bucket.
-    static constexpr auto kTimeseriesBucketMaxTimeRange = Hours(1);
-
     class BucketId {
         friend class BucketCatalog;
 
diff --git a/src/mongo/db/timeseries/timeseries.idl b/src/mongo/db/timeseries/timeseries.idl
index 048a51f4e93..c8c65dd3e65 100644
--- a/src/mongo/db/timeseries/timeseries.idl
+++ b/src/mongo/db/timeseries/timeseries.idl
@@ -61,6 +61,16 @@ server_parameters:
         cpp_varname: gTimeseriesBucketsCollectionClusterById
         default: true
 
+enums:
+    BucketGranularity:
+        description: "Describes a time-series collection's expected interval between subsequent
+                      measurements"
+        type: string
+        values:
+            Seconds: "seconds"
+            Minutes: "minutes"
+            Hours: "hours"
+
 structs:
     TimeseriesOptions:
         description: "The options that define a time-series collection."
@@ -82,3 +92,12 @@ structs:
                               deleted."
                 type: safeInt64
                 optional: true
+            granularity:
+                description: "Describes the expected interval between subsequent measurements"
+                type: BucketGranularity
+                default: Seconds
+            bucketMaxSpanSeconds:
+                description: "The maximum range of time values for a bucket, in seconds"
+                type: safeInt
+                default: 3600
+                validator: { gte: 1 }
diff --git a/src/mongo/db/ttl.cpp b/src/mongo/db/ttl.cpp
index 68c7ed8a294..18f84480343 100644
--- a/src/mongo/db/ttl.cpp
+++ b/src/mongo/db/ttl.cpp
@@ -57,6 +57,7 @@
 #include "mongo/db/timeseries/bucket_catalog.h"
 #include "mongo/db/ttl_collection_cache.h"
 #include "mongo/db/ttl_gen.h"
+#include "mongo/db/views/view_catalog.h"
 #include "mongo/logv2/log.h"
 #include "mongo/util/background.h"
 #include "mongo/util/concurrency/idle_thread_block.h"
@@ -294,15 +295,27 @@ private:
      * Generate the safe expiration date for a given collection and user-configured
      * expireAfterSeconds value.
      */
-    Date_t safeExpirationDate(const CollectionPtr& coll, std::int64_t expireAfterSeconds) const {
+    Date_t safeExpirationDate(OperationContext* opCtx,
+                              const CollectionPtr& coll,
+                              std::int64_t expireAfterSeconds) const {
         if (coll->ns().isTimeseriesBucketsCollection()) {
-            // Don't delete data unless it is safely out of range of the bucket maximum time range.
-            // On time-series collections, the _id (and thus RecordId) is the minimum time value of
-            // a bucket. A bucket may have newer data, so we cannot safely delete the entire bucket
-            // yet until the maximum bucket range has passed, even if the minimum value can be
-            // expired.
-            return Date_t::now() - Seconds(expireAfterSeconds) -
-                Seconds(BucketCatalog::kTimeseriesBucketMaxTimeRange);
+            auto timeseriesNs = coll->ns().bucketsNamespaceToTimeseries();
+            auto viewCatalog = DatabaseHolder::get(opCtx)->getViewCatalog(opCtx, timeseriesNs.db());
+            invariant(viewCatalog);
+            auto viewDef = viewCatalog->lookup(opCtx, timeseriesNs.ns());
+            uassert(ErrorCodes::NamespaceNotFound,
+                    fmt::format("Could not find view definition for namespace: {}",
+                                timeseriesNs.toString()),
+                    viewDef);
+
+            const auto bucketMaxSpan = Seconds(viewDef->timeseries()->getBucketMaxSpanSeconds());
+
+            // Don't delete data unless it is safely out of range of the bucket maximum time
+            // range. On time-series collections, the _id (and thus RecordId) is the minimum
+            // time value of a bucket. A bucket may have newer data, so we cannot safely delete
+            // the entire bucket yet until the maximum bucket range has passed, even if the
+            // minimum value can be expired.
+            return Date_t::now() - Seconds(expireAfterSeconds) - bucketMaxSpan;
         }
 
         return Date_t::now() - Seconds(expireAfterSeconds);
@@ -375,7 +388,8 @@ private:
 
         const Date_t kDawnOfTime =
             Date_t::fromMillisSinceEpoch(std::numeric_limits<long long>::min());
-        const auto expirationDate = safeExpirationDate(collection, secondsExpireElt.numberLong());
+        const auto expirationDate =
+            safeExpirationDate(opCtx, collection, secondsExpireElt.numberLong());
         const BSONObj startKey = BSON("" << kDawnOfTime);
         const BSONObj endKey = BSON("" << expirationDate);
         // The canonical check as to whether a key pattern element is "ascending" or
@@ -459,7 +473,7 @@ private:
                     "running TTL job for collection clustered by _id",
                     logAttrs(collection->ns()));
 
-        const auto expirationDate = safeExpirationDate(collection, *expireAfterSeconds);
+        const auto expirationDate = safeExpirationDate(opCtx, collection, *expireAfterSeconds);
 
         // Generate upper bound ObjectId that compares greater than every ObjectId with a the same
         // timestamp or lower.