Show how to fetch logs from long-term HDFS storage

01a846b4 · Konstantinos Samaras-Tsakiris · 5e0a0521 · 01a846b4 · 01a846b4 · 01a846b4
Commit 01a846b4 authored 2 years ago by Konstantinos Samaras-Tsakiris
--- a/README.md
+++ b/README.md
@@ -3,3 +3,7 @@
 This project includes operations on Drupal sites or across infrastructure that are not part of the drupalSite-operator, such as Tekton tasks.

 We implement actions that the infrastructure users can apply ad-hoc to their websites, and also other infrastructure components can use to perform their tasks.
+
+## Examples
+
+In Examples there are some one-off operations that show how something could be done, without providing full automation.
--- a/examples/README.md
+++ b/examples/README.md
+### tekton-tasks
+
+Examples of backing up or restoring Drupal sites.
+
+### logs-hdfs
+
+How to fetch site logs from long-term storage on HDFS.
+
+This Jupyter notebook should be run on the SWAN service ([swan.cern.ch](https://swan.cern.ch))
+using the SPARK plugin.
--- a/examples/logs-hdfs/fetchHomeData.ipynb
+++ b/examples/logs-hdfs/fetchHomeData.ipynb
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "bb5bd38c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from pyspark.sql.functions import col, asc, countDistinct, date_format, from_unixtime\n",
+    "from pyspark.sql import functions as F\n",
+    "from datetime import date, timedelta\n",
+    "import pandas as pd\n",
+    "from pyspark.sql import DataFrame"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "05d5f738",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "start_date = date(2020, 8, 1)\n",
+    "duration= timedelta(days=31)\n",
+    "\n",
+    "def fetchLogs(dates):\n",
+    "    paths= ['/project/monitoring/archive/drupal/logs/prod8/drupal_service/drupal8/'+date.strftime(\"%Y/%m/%d\")+'/*'\n",
+    "            for date in dates]\n",
+    "    return spark.read.json(paths)\n",
+    "    #return sc.union(spark.read.json(paths))\n",
+    "def selectClientip(log, whereFilt):\n",
+    "    return log.select(col(\"data.clientip\"), date_format(from_unixtime(col(\"metadata.timestamp\")/1000), \"yyyy-MM-dd\")\n",
+    "                       .alias(\"timestamp\")).where(whereFilt)\n",
+    "def concatClientipLogs(start_date, duration, whereFilt):\n",
+    "    dates= [start_date + timedelta(days=d) for d in range(duration.days+1)]\n",
+    "    return selectClientip(fetchLogs(dates), whereFilt)\n",
+    "\n",
+    "#https://spark.apache.org/docs/latest/api/python/reference/api/pyspark.sql.functions.transform.html#pyspark.sql.functions.transform\n",
+    "def datebinning(col):\n",
+    "    #return F.dayofyear(col)\n",
+    "    #return F.weekofyear(col)\n",
+    "    return F.month(col)\n",
+    "def countUniqueIPinDatebin(df):\n",
+    "    return df.groupBy(\"datebin\").agg(countDistinct(\"clientip\"))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "24fd5729",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "homeAccessLogs = concatClientipLogs(start_date, duration, 'data.program == \"httpd\" AND data.sitename == \"home.cern\"')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "63c5d412",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "homeAccessLogsBinned= countUniqueIPinDatebin(homeAccessLogs.withColumn('datebin', datebinning('timestamp')).select([\"clientip\",\"datebin\"])).toPandas().sort_values(by=\"datebin\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "ed7b9057",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>datebin</th>\n",
+       "      <th>count(clientip)</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>8</td>\n",
+       "      <td>415728</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>9</td>\n",
+       "      <td>17528</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "   datebin  count(clientip)\n",
+       "1        8           415728\n",
+       "0        9            17528"
+      ]
+     },
+     "execution_count": 5,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "homeAccessLogsBinned.sort_values(by=\"datebin\")\n",
+    "# F.month:\n",
+    "# datebin\tcount(clientip)\n",
+    "#0\t12\t108026\n",
+    "#1\t9\t334389\n",
+    "#2\t10\t306606\n",
+    "#3\t11\t335701\n",
+    "\n",
+    "# F.dayofyear:\n",
+    "# avg(Sept): 13291\n",
+    "# 0.8386309601436553"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "2e3536a1",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "homeAccessLogsBinned.sort_values(by=\"datebin\").to_csv(\"homeAccessLogs.csv\", index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "3fc7d98b",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.9.6"
+  },
+  "sparkconnect": {
+   "bundled_options": [],
+   "list_of_options": []
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}
+%% Cell type:code id:bb5bd38c tags:
+
+``` python
+from pyspark.sql.functions import col, asc, countDistinct, date_format, from_unixtime
+from pyspark.sql import functions as F
+from datetime import date, timedelta
+import pandas as pd
+from pyspark.sql import DataFrame
+```
+
+%% Cell type:code id:05d5f738 tags:
+
+``` python
+start_date = date(2020, 8, 1)
+duration= timedelta(days=31)
+
+def fetchLogs(dates):
+    paths= ['/project/monitoring/archive/drupal/logs/prod8/drupal_service/drupal8/'+date.strftime("%Y/%m/%d")+'/*'
+            for date in dates]
+    return spark.read.json(paths)
+    #return sc.union(spark.read.json(paths))
+def selectClientip(log, whereFilt):
+    return log.select(col("data.clientip"), date_format(from_unixtime(col("metadata.timestamp")/1000), "yyyy-MM-dd")
+                       .alias("timestamp")).where(whereFilt)
+def concatClientipLogs(start_date, duration, whereFilt):
+    dates= [start_date + timedelta(days=d) for d in range(duration.days+1)]
+    return selectClientip(fetchLogs(dates), whereFilt)
+
+#https://spark.apache.org/docs/latest/api/python/reference/api/pyspark.sql.functions.transform.html#pyspark.sql.functions.transform
+def datebinning(col):
+    #return F.dayofyear(col)
+    #return F.weekofyear(col)
+    return F.month(col)
+def countUniqueIPinDatebin(df):
+    return df.groupBy("datebin").agg(countDistinct("clientip"))
+```
+
+%% Cell type:code id:24fd5729 tags:
+
+``` python
+homeAccessLogs = concatClientipLogs(start_date, duration, 'data.program == "httpd" AND data.sitename == "home.cern"')
+```
+
+%% Cell type:code id:63c5d412 tags:
+
+``` python
+homeAccessLogsBinned= countUniqueIPinDatebin(homeAccessLogs.withColumn('datebin', datebinning('timestamp')).select(["clientip","datebin"])).toPandas().sort_values(by="datebin")
+```
+
+%% Cell type:code id:ed7b9057 tags:
+
+``` python
+homeAccessLogsBinned.sort_values(by="datebin")
+# F.month:
+# datebin	count(clientip)
+#0	12	108026
+#1	9	334389
+#2	10	306606
+#3	11	335701
+
+# F.dayofyear:
+# avg(Sept): 13291
+# 0.8386309601436553
+```
+
+%% Output
+
+       datebin  count(clientip)
+    1        8           415728
+    0        9            17528
+
+%% Cell type:code id:2e3536a1 tags:
+
+``` python
+homeAccessLogsBinned.sort_values(by="datebin").to_csv("homeAccessLogs.csv", index=False)
+```
+
+%% Cell type:code id:3fc7d98b tags:
+
+``` python
+```
--- a/examples/clear-cache-taskrun.yaml
+++ b/examples/clear-cache-taskrun.yaml
--- a/examples/database-restore-taskrun.yaml
+++ b/examples/database-restore-taskrun.yaml
--- a/examples/drupalsite-backup-taskrun.yaml
+++ b/examples/drupalsite-backup-taskrun.yaml
--- a/examples/drupalsite-restore-taskrun.yaml
+++ b/examples/drupalsite-restore-taskrun.yaml
@@ -8,9 +8,9 @@ spec:
    kind: ClusterTask
  params:
    - name: drupalSite
-      value: drupalsite-sample
+      value: arts
    - name: backupName
-      value: ravineet-1-tekton-test-fbfe0
+      value: arts-901a-20220530000645
    - name: namespace
-      value: ravineet-1
+      value: arts
  serviceAccountName: tektoncd