usage monitoring app

Streamlit in Snowflake Usage Monitoring/README.md

@@ -0,0 +1,79 @@
+![](../shared_assets/sis-header.jpeg)
+
+# Streamlit in Snowflake Usage Monitoring
+
+This app helps Accountadmins monitor the usage of Streamlit in Snowflake apps. It uses the [Account Usage](https://docs.snowflake.com/en/sql-reference/account-usage) schema to query historical usage data for your account. 
+
+
+## Prerequisites
+
+The App Owner Role must have SELECT privileges on the [Query History view](https://docs.snowflake.com/en/sql-reference/account-usage/query_history). 
+
+
+## App Layout
+
+This app contains 2 tabs:
+
+1. Account Level App Usage
+2. App Level Usage
+
+
+### Account Level Usage
+
+![](./assets/account_level.png)
+
+### App Level Usage
+
+![](./assets/app_level.png)
+
+### Raw Data
+
+The app also shows the Raw data powering the graphs above. 
+
+
+## Streamlit Execution Metadata
+
+All Streamlit in Snowflake apps executions are recorded in the [Query History view](https://docs.snowflake.com/en/sql-reference/account-usage/query_history). This can be visualized by going to `Monitoring` -> `Query History` in Snowsight. For example, all Streamlit app executions show up as 
+
+```sql
+execute streamlit "ST_TEST_DB"."DEV"."KP1V05SACRMOJ9PX"()
+```
+
+You can also run a SQL query in a Worksheet to obtain the same information. 
+
+```sql
+select 
+    user_name as User,
+    total_elapsed_time/60000 as Minutes_Used, 
+    date(start_time) as Date,
+    try_parse_json(query_tag):StreamlitName as AppName,
+    query_text
+from snowflake.monitoring.query_history
+where 
+    -- Enter a date in the line below
+    start_time >= '{start_date}'
+    and try_parse_json(query_tag):StreamlitEngine = 'ExecuteStreamlit'
+    and try_parse_json(query_tag):ChildQuery IS NULL
+    and contains(query_text, 'execute streamlit') 
+    order by Date desc;
+```
+
+All Streamlit in Snowflake app executions contain a `QUERY_TAG`, for e.g. 
+
+```json
+{
+    "StreamlitEngine":"ExecuteStreamlit",
+    "StreamlitName":"ST_TEST_DB.DEV.KP1V05SACRMOJ9PX"
+}
+```
+
+Any child queries executed by the app also contain an additional field in the `QUERY_TAG` indicating that it is a child query
+
+```json
+{
+    "StreamlitEngine":"ExecuteStreamlit",
+    "StreamlitName":"ST_TEST_DB.DEV.KP1V05SACRMOJ9PX",
+    "ChildQuery":true
+}
+```
+

Streamlit in Snowflake Usage Monitoring/pages/app_level_metrics.py

@@ -0,0 +1,117 @@
+# Import python packages
+import streamlit as st
+from snowflake.snowpark.context import get_active_session
+import datetime 
+import altair as alt
+import pandas as pd
+
+
+today = datetime.date.today()
+yesterday = today - datetime.timedelta(days=1)
+week_ago = today - datetime.timedelta(days=7)
+last_30_days = today - datetime.timedelta(days=30)
+
+
+# Write directly to the app
+st.title("App Level Usage Monitoring :balloon:")
+st.write(
+    """Metrics about the usage of a specific app in your account
+    """
+)
+
+def generate_sql(start_date):
+    sql = f"""
+    select 
+        user_name as User,
+        total_elapsed_time/60000 as Minutes_Used, 
+        date(start_time) as Date,
+        try_parse_json(query_tag):StreamlitName as AppName,
+        query_text
+    from snowflake.monitoring.query_history
+    where start_time >= '{start_date}'
+    and try_parse_json(query_tag):StreamlitEngine = 'ExecuteStreamlit'
+    and try_parse_json(query_tag):ChildQuery IS NULL
+    and contains(query_text, 'execute streamlit') 
+    order by Date desc;
+    """
+    return sql
+
+@st.cache_data
+def query_sf(sql):
+    df = session.sql(sql).to_pandas()
+    return df
+
+
+def generate_bar_chart(views_over_time):
+    views_over_time.index = pd.to_datetime(views_over_time.index)  # Convert index to datetime
+
+    # Convert Series to DataFrame for Altair
+    df = views_over_time.reset_index()
+    df.columns = ['Date', 'Value']  # Rename columns for clarity
+
+    # Create bar chart
+    chart = alt.Chart(df).mark_bar().encode(
+        x=alt.X('Date:T', title='Date'),
+        y=alt.Y('Value:Q', title='Number of Views'),
+        tooltip=['Date', 'Value']  # Add tooltips to display the date and value on hover
+    ).properties(
+    width=600,
+    height=400
+    )
+
+    st.altair_chart(chart)    
+
+
+@st.cache_data
+def calculate_metrics(df):
+    num_users = df['USER'].nunique()
+    num_apps = df['APPNAME'].nunique()
+    num_views = len(df.index) 
+
+    st.subheader('Summary Stats')
+    col1, col2, col3 = st.columns(3)
+
+    with col1:
+        st.metric(label='Unique Users', value=num_users)
+
+    with col2:
+        st.metric(label='Apps Used', value=num_apps)
+
+    with col3:
+        st.metric(label='Total Views', value=num_views)
+
+    st.subheader('Views over time')
+    views_over_time = df.groupby(['DATE']).size()
+    generate_bar_chart(views_over_time)
+
+    st.subheader('Raw Data')
+    st.dataframe(df)
+
+session = get_active_session()
+
+sql = generate_sql(last_30_days)
+df = query_sf(sql)
+list_of_apps = df['APPNAME'].unique()
+
+app = st.selectbox('Choose App Name', list_of_apps)
+
+tab_yesterday, tab_last_week, tab_last_30_days = st.tabs(["Last Day", "Last Week", "Last 30 Days"])
+
+with tab_yesterday:
+    sql = generate_sql(yesterday)
+    df = query_sf(sql)
+    df2 = df[df['APPNAME'] == app]
+    calculate_metrics(df2)
+
+
+with tab_last_week:
+    sql = generate_sql(week_ago)
+    df = query_sf(sql)
+    df2 = df[df['APPNAME'] == app]
+    calculate_metrics(df2)
+
+with tab_last_30_days:
+    sql = generate_sql(last_30_days)
+    df = query_sf(sql)
+    df2 = df[df['APPNAME'] == app]
+    calculate_metrics(df2)

Streamlit in Snowflake Usage Monitoring/streamlit_app.py

@@ -0,0 +1,108 @@
+# Import python packages
+import streamlit as st
+from snowflake.snowpark.context import get_active_session
+import datetime 
+import altair as alt
+import pandas as pd
+
+
+today = datetime.date.today()
+yesterday = today - datetime.timedelta(days=1)
+week_ago = today - datetime.timedelta(days=7)
+last_30_days = today - datetime.timedelta(days=30)
+
+
+# Write directly to the app
+st.title("Streamlit in Snowflake Usage :balloon:")
+st.write(
+    """Metrics about the overall Streamlit in Snowflake usage in your account
+    """
+)
+
+def generate_sql(start_date):
+    sql = f"""
+    select 
+        user_name as User,
+        total_elapsed_time/60000 as Minutes_Used, 
+        date(start_time) as Date,
+        try_parse_json(query_tag):StreamlitName as AppName,
+        query_text
+    from snowflake.monitoring.query_history
+    where start_time >= '{start_date}'
+    and try_parse_json(query_tag):StreamlitEngine = 'ExecuteStreamlit'
+    and try_parse_json(query_tag):ChildQuery IS NULL
+    and contains(query_text, 'execute streamlit') 
+    order by Date desc;
+    """
+    return sql
+
+@st.cache_data
+def query_sf(sql):
+    df = session.sql(sql).to_pandas()
+    return df
+
+
+def generate_bar_chart(views_over_time):
+    views_over_time.index = pd.to_datetime(views_over_time.index)  # Convert index to datetime
+
+    # Convert Series to DataFrame for Altair
+    df = views_over_time.reset_index()
+    df.columns = ['Date', 'Value']  # Rename columns for clarity
+
+    # Create bar chart
+    chart = alt.Chart(df).mark_bar().encode(
+        x=alt.X('Date:T', title='Date'),
+        y=alt.Y('Value:Q', title='Number of Views'),
+        tooltip=['Date', 'Value']  # Add tooltips to display the date and value on hover
+    ).properties(
+    width=600,
+    height=400
+    )
+
+    st.altair_chart(chart)    
+
+
+@st.cache_data
+def calculate_metrics(df):
+    num_users = df['USER'].nunique()
+    num_apps = df['APPNAME'].nunique()
+    num_views = len(df.index) 
+
+    st.subheader('Summary Stats')
+    col1, col2, col3 = st.columns(3)
+
+    with col1:
+        st.metric(label='Unique Users', value=num_users)
+
+    with col2:
+        st.metric(label='Apps Used', value=num_apps)
+
+    with col3:
+        st.metric(label='Total Views', value=num_views)
+
+    st.subheader('Views over time')
+    views_over_time = df.groupby(['DATE']).size()
+    generate_bar_chart(views_over_time)
+
+    st.subheader('Raw Data')
+    st.dataframe(df)
+
+session = get_active_session()
+
+tab_yesterday, tab_last_week, tab_last_30_days = st.tabs(["Last Day", "Last Week", "Last 30 Days"])
+
+with tab_yesterday:
+    sql = generate_sql(yesterday)
+    df = query_sf(sql)
+    calculate_metrics(df)
+
+
+with tab_last_week:
+    sql = generate_sql(week_ago)
+    df = query_sf(sql)
+    calculate_metrics(df)
+
+with tab_last_30_days:
+    sql = generate_sql(last_30_days)
+    df = query_sf(sql)
+    calculate_metrics(df)